\ PH0^S°* OFFICE OF INSPECTOR GENERAL
Catalyst for Improving the Environment
Evaluation Report
EPA Needs to Direct More Attention,
Efforts, and Funding to Enhance Its
Speciation Monitoring Program for
Measuring Fine Particulate Matter
Report No. 2005-P-00004
February 7, 2005

-------
Report Contributors:
Patrick Milligan
Mark S. Phillips
Kevin Good
Sarah Fabirkiewicz
Abbreviations
APM	Annual Performance Measure
CEM	Continuous Emissions Monitor
EPA	Environmental Protection Agency
FRM	Federal Reference Method
GPRA	Government Performance and Results Act
IMPROVE	Interagency Monitoring of PROtected Visual Environments
NAAQS	National Ambient Air Quality Standards
NARSTO	North American Research Strategy for Tropospheric Ozone
NEI	National Emissions Inventory
NOx	Nitrogen Oxide
OAQPS	Office of Air Quality Planning and Standards
OAR	Office of Air and Radiation
OIG	Office of Inspector General
ORD	Office of Research and Development
PM2 5	Particulate Matter2 5
RPO	Regional Planning Organization
SIP	State Implementation Plan
S02	Sulfur Dioxide
STAR	Science to Achieve Results
STN	Speciation T rends N etwork
Cover Photo:
Fine particles come from a variety of sources.
Source: US EPA, Region 1, presentation titled Fine Particles in the Air
dated April 10, 2003.

-------
^tosr/iv
2
<
5o
\.
'u
t-
o
*
*¦> /
PflCA^0
.<*
*?
U.S. Environmental Protection Agency
Office of Inspector General
At a Glance
2005-P-00004
February 7, 2005
Why We Did This Review
We sought to determine
whether EPA's PM2 5
speciation air monitoring
network is sufficient to
(a) adequately identify sources
of fine particulate matter
(PM2 5), and (b) facilitate the
development of effective
control strategies to reduce
PM2 5 to safe levels.
Determining the chemical
make-up of a particle - known
as speciation - is largely
accomplished through data
generated by this network.
Background
Airborne particulate matter
2.5 microns or less in size
(PM2 5) is comprised of a
complex mixture of particles
composed of sulfate, nitrate,
ammonium, organic carbon,
elemental carbon, and organic
and inorganic compounds.
Tens of thousands of
premature deaths yearly are
associated with exposure to
excess levels of PM2 5. By
2010, EPA estimates that
compliance with PM2 5
emission control strategies
will cost industry more than
$37 billion annually. EPA's
speciation monitoring network
is a critical component in the
development of these control
strategies.
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.gov/oig/reports/2005/
20050207-2005-P-00004
Catalyst for Improving the Environment
EPA Needs to Direct More Attention, Efforts, and Funding
to Enhance Its Speciation Monitoring Program for Measuring
Fine Particulate Matter
What We Found
EPA has made substantial progress in establishing a speciation monitoring
network to facilitate the development of PM2 5 control strategies, but still faces a
number of challenges in ensuring that the controls are implemented at the right
sources. The development of control strategies is best approached through
collaborative processes that use emissions inventories, ambient monitoring data,
and air quality modeling. Although the speciation network provides information
for understanding the make-up and origin of PM2 5, the network does not fully
assist in providing the data for EPA and States to identify or quantify the chemical
make-up of PM2 5 particles, reliably trace particles back to their source, or account
for chemical changes that occur after particles are released into the atmosphere.
Speciation data are available to begin working on control strategies, and EPA and
the States are beginning the development of control strategies; however, increased
monitoring efforts are needed.
Under the Clean Air Act, States with PM2 5 nonattainment areas have until
February 2008 to develop control strategies for reducing PM2 5, and an additional
2 years to reach attainment with the PM2 5 standard. Also, with justification, the
Act allows EPA to grant a State an extension of up to 5 years to reach full
attainment. Data from EPA's speciation network will be vital to ensuring that
pollution controls are implemented at the right sources. Otherwise, some
facilities may install unneeded controls, while some needed controls may go
uninstalled; ultimately, compliance may be further delayed and more costly.
Agency officials acknowledge that improved speciation data will be needed for
EPA to overcome the uncertainties associated with PM2 5 particle origin. In 2004,
EPA budgeted over $43 million for PM2 5 monitoring, with about $16.4 million
for operation of the existing speciation monitoring network. However, only about
$800,000 was budgeted for improving its capability to address uncertainties with
PM2 5 particle origin. According to manufacturers and some Agency officials we
contacted, increased partnering between EPA and monitor manufacturers may be
needed if advanced speciation monitors are to be developed in time to help
agencies develop air pollution control strategies that ensure controls are
implemented at the right sources.
What We Recommend
We recommend that EPA increase its research on technologies that can more fully
identify the chemical make-up of PM2 5, account for the atmospheric impacts on
PM2 5, and assay the resultant changes that occur to the composition of the
particle. This includes increasing opportunities for cooperation with the private
sector to develop improved continuous speciation monitors. In its response to the
draft report, EPA disagreed with certain issues; however, the Agency stated that
the recommendations generally align with their current improvement efforts.

-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
OFFICE OF
INSPECTOR GENERAL
February 7, 2005
MEMORANDUM
SUBJECT:	EPA Needs to Direct More Attention, Efforts, and Funding to Enhance Its
Speciation Monitoring Program for Measuring Fine Particulate Matter
Report No. 2005-P-00004
FROM:	J. Rick Beusse /s/
Director for Program Evaluation, Air Issues
TO:	Jeffrey R. Holmstead
Assistant Administrator for Air and Radiation (6101 A)
William Farland
Acting Deputy Assistant Administrator for Science,
Office of Research and Development (8101R)
Attached is our final report regarding the Environmental Protection Agency (EPA) Particulate
Matter Ambient Speciation Monitoring Program. This report contains findings regarding EPA's
need to direct more attention, effort, and funding toward its Ambient Speciation Monitoring
Program. Also, the report contains corrective actions the Office of Inspector General (OIG)
recommends. This report represents the opinion of the OIG, and the findings contained in this
report do not necessarily represent the final EPA position. Final determination on matters in this
report will be made by EPA managers in accordance with established procedures.
EPA's Office of Air and Radiation provided us with a response on January 31, 2005, that
consolidated its comments to the draft report with those from the Office of Research and
Development. We included EPA's consolidated response in its entirety as Appendix H.
Action Required
In accordance with EPA Manual 2750, as the action official, you are required to provide this office
with a written response within 90 days of the final report date. Since this report deals primarily
with the EPA Office of Air and Radiation's Ambient Speciation Program, the Assistant
Administrator for Air and Radiation was designated the primary action official. As such, he should
take the lead in coordinating the Agency's response. 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. If you do not
concur with a recommendation, please provide alternative actions addressing the findings reported.
We appreciate the efforts of EPA officials and staff, as well as external stakeholders, in working
with us to develop this report. For your convenience, this report will be available at
http ://www.epa.gov/ oig.
If you or your staff have any questions regarding this report, please contact me at (919) 541-5747
or Patrick Milligan, Assignment Manager, at (215) 814-2326.

-------
Table of C
At a Glance
Chapters
1	Introduction 		1
Purpose 		1
Background		2
Scope and Methodology 		3
Results in Brief		5
2	Speciation Monitoring Critical to Controlling
Fine Particulate Matter and Reaching Attainment		7
Nonattainment Areas Recommended 		7
Speciation Data Critical to Overall Success of PM25 Program		11
Speciation Data Used to Groundtruth Emissions
Inventory Estimates and Modeling Assumptions 		13
Speciation Data Needed to Improve Understanding of
PM Exposure and Health Effects		16
3	EPA Faces Many Challenges in Identifying
and Controlling Fine Particulate Matter 	17
Limitations of Current Tools Used to Monitor and
Assess PM25 Air Quality		17
Challenges in Measuring Carbon and Ammonium, and
Accounting for Transport, Make It Difficult to Fully
Identify and Quantify Components of PM2 5 		18
EPA Has Made Efforts to Address Challenges in Obtaining
Speciation Data to Identify Sources and Develop Control Strategies ....	23
Increased Partnering with Monitor Manufacturers Could
Improve Monitoring Capabilities and Uses		29
Conclusions		30
Recommendations		31
Agency Comments and OIG Evaluation		33

-------
Appendices
A Major Steps in SIP Development and Approval Process 		34
B National Emission Control Programs Providing Benefit to
PM25 Emission Reduction Efforts		36
C Description of the Two Primary Ambient Speciation Networks		38
D PM25 Emissions Inventory and Its Relationship to
Monitoring and Modeling 		39
E Atmospheric Modeling and Its Relationship to
Monitoring and Emissions Inventory		40
F Advanced PM25 Research Conducted by Supersites Program		43
G Recent Grant Awards from ORD's STAR Solicitation		45
H Consolidated EPA Response to Draft Report 		46
I OIG Evaluation of Consolidated EPA Response to Draft Report		57
J Distribution		60

-------
Chapter 1
Introduction
Purpose
Severe health effects are associated with exposure to excess levels of airborne
fine particulate matter (PM2 5), including tens of thousands of premature deaths and
hospital admissions, and hundreds of thousands of doctor visits, work and school
absences, and respiratory illnesses yearly. The Environmental Protection Agency
(EPA) first established the PM2 5 standard1 in 1997, and in September 2004
reconfirmed the serious health effects from exposure to excess levels of PM25. In
June 2004, EPA alerted 21 States that 244 counties, with a collective population of
99 million people, were likely to exceed the PM2 5 standard.
Determining the chemical make-up of a particle - known as "speciation" - is an
important part of the effort to reduce PM2 5 levels, and is accomplished largely
through data generated by EPA's ambient air speciation monitoring program. The
program is designed to help EPA and the States better understand the chemical
composition of the particle; what happens to the particle after it is released in the
atmosphere; how the particle can be traced to its source of origin, also known as
source apportionment; whether implemented controls are having the desired effect
on air quality; and the potential health effects of PM2 5. To reduce ambient air levels
sufficient to attain the PM2 5 standard, EPA, State, local, and tribal agencies will
have to overcome substantial challenges in identifying and controlling the sources
of PM2 5. EPA's speciation monitoring is central to identifying those sources,
facilitating the development of effective control strategies, and gauging their
success. EPA estimates that compliance with PM2 5 emission control strategies will
cost industry more than $37 billion annually by 2010. Thus, it is critical that
controls be implemented at the right sources. Otherwise, some facilities may install
unneeded controls, while some needed controls may go uninstalled; ultimately,
compliance may be further delayed and more costly.
EPA recognizes the importance of speciation data to achieving its long-term goals,
and has efforts underway to improve its speciation monitoring program. Therefore,
we examined the challenges facing EPA with the intent of bringing attention to
areas on which the Agency should further focus its efforts. Specifically, we sought
to determine whether the PM2 5 ambient air speciation monitoring program is
sufficient to: (1) adequately identify sources of PM2 5, and (2) facilitate the
development of effective control strategies to reduce PM2 5 to safe levels.
'EPA's PM, 5 standard requires that levels of fine particles remain at or below 65 micrograms per cubic
meter of air over a 24-hour period, and at or below 15 micrograms per cubic meter of air on an average annual basis.
1

-------
Background
PM2 5 and Its Health Effects
Particulate Matter (PM) includes acids, metals, and the solid or liquid droplets in
gases, and other harmful airborne substances that can be breathed into the lungs.
PM sources include automobiles, diesel engines, power plants, industrial facilities,
wood combustion, and dust from roads. Some particles are large or dark enough to
be seen as soot or smoke; others can only be detected with an electron microscope.
EPA has been concerned about the adverse effects of PM on human health and the
environment since the early 1970s. The first airborne particles to be regulated were
referred to as Total Suspended Particulate Matter, which included a broad range of
large and small particles. Today, EPA no longer monitors for Total Suspended
Particulate Matter, but instead regulates several smaller-sized particles, since they
are more likely to slip past body defenses (nose, throat, and larynx) and penetrate
deep into the lungs. EPA regulates two types of smaller airborne particles, as
shown in Table 1.1.
Table 1.1: Types of Regulated Particulate Matter
Type8
Description
Date Regulated
PM10
Particles less than or equal to 10 microns in diameter
(about one-seventh the diameter of a human hair).
1987
pm2.5
"Fine" particles, which are less than or equal to 2.5 microns
in diameter (about 1/30th the diameter of a human hair).
1997
aA new PM standard - PMC (known as "coarse") - is being considered by EPA to apply to the fraction of PM
between 2.5 and 10 microns. EPA's current schedule should provide for a final standard in late 2005.
Every 5 years, EPA revisits standards to ensure they reflect current information and
are protective of human health. The newer category - PM2 5 - was established as a
National Ambient Air Quality Standard (NAAQS) in 1997. This standard requires
that levels of fine particles remain at or below 65 micrograms per cubic meter of
air over a 24-hour period, and at or below 15 micrograms per cubic meter of air on
an average annual basis. EPA established this standard as a result of a growing
body of scientific evidence indicating that these fine particles are most damaging to
health since they can penetrate the lung tissues easier and deeper.
Compliance with NAAQS is measured using ambient monitoring, but because
people experience adverse health effects from the air that they breathe, it is
important to understand how ambient concentrations relate to actual human
exposures. For PM2 5 recent exposure studies have demonstrated that ambient
monitors are reasonable surrogates for human exposure to ambient PM2 5 and
sulfates. However, for other components of ambient PM, relationships between
ambient concentrations and actual human exposures have not been established.
2

-------
When breathed, particulate matter can accumulate in the respiratory system. Fine
particulate matter is associated with such adverse health effects as heart and lung
disease and increased respiratory disease, and symptoms such as asthma, decreased
lung function, and even premature death. Sensitive groups that appear to be at
greatest risk include the elderly, individuals with cardiopulmonary disease, and
children. Also, PM is a major cause of reduced visibility, and adversely impacts
vegetation and ecosystems.
PM2 5 in the atmosphere is composed of a complex mixture of particles: sulfate,
nitrate, and ammonium2 particles; organic carbon composed of a large number of
individual organic species and elemental carbon; and other inorganic material.
•	"Primary" particles are emitted directly into the air as solid or liquid particles.
Examples include elemental carbon from diesel engines or forest fires, and
condensible organic particles from gasoline engines.
•	"Secondary" fine particles are formed in the atmosphere through the chemical
reactions of precursor gas emissions, including organic gases, nitrogen oxides,
sulfur oxides, and ammonia. Often at least half of the PM2 5 mass consists of
secondary particles formed through a change in composition, making it a
challenge to identify a particle's source of origin.
Furthermore, depending on particle size and meteorological conditions, such as
wind speed and direction, excess levels of PM2 5 and precursor species that
originated in one area maybe transported by the wind to another area. Fine
particles below 2 microns may travel thousands of miles, while larger particles - 10
microns in size or larger - may only travel a few hundred meters. For example,
certain eastern States have alleged that some of their particulate problems can be
attributed to power plants located hundreds of miles away in the Upper Ohio
Valley.
Scope and Methodology
To assess whether the PM2 5 speciation monitoring network is sufficient to identify
sources of PM2 5 and facilitate the development of effective control strategies to
reduce PM2 5by State, local, and tribal agencies, we discussed PM2 5 speciation data
challenges, monitoring capabilities, and monitoring limitations with officials from:
•	EPA's Office of Air Quality Planning and Standards
2
Two common forms of secondarily formed PM25 occur when acid sulfates and nitric acid react with
ammonia in the atmosphere, creating ammonium sulfate and ammonium nitrate, respectively. Ammonium (NH4) is
made up of ammonia (NH3) and hydrogen (H), and is formed when NH3 gas reacts with a hydrogen ion from an
acidic species either in the gas phase or in solution (e.g., NH3(g) + HN03(g) = NH4N03(s)). Ammonium is only in
the aerosol phase, while NH3 is in the gas phase or can be dissolved in water where it quickly reacts with an acidic
species.
3

-------
•	EPA's Office of Research and Development
•	EPA Regions 3,5, and 9
•	Selected State air pollution control agencies (Pennsylvania Department of
Environmental Protection; California Air Resources Board; Illinois
Environmental Protection Agency; North Carolina Department of
Environment, Health, and Natural Resources, Air Protection Division; and the
Georgia Department of Natural Resources, Environmental Protection Division)
•	The State and Territorial Air Pollution Program Administrators/Association of
Local Air Pollution Control Officials
•	Regional Planning Organizations (the Mid-Atlantic Regional Air Management
Association,3 and the Central Regional Air Planning Association4
•	Three major PM2 5 monitor manufacturers (Met-One, Inc., Rupprecht &
Patashnik Co., Inc., and Thermo, Inc.)
•	Academia (Clarkson University and University of Maryland)
We also reviewed key reports and studies related to PM2 5 and specifically the
speciation of PM2 5, including:
•	NARSTO report - Particulate Matter Science for Policymakers: A NARSTO5
Assessment
•	National Research Council's report - Research Priorities for Airborne
Particulate Matter, Volume IV
•	National Research Council's report - Air Quality Management of the United
States
•	Clean Air Interstate Rule
•	MANE-VU's Speciation Network Data Review
•	Draft of 40 Code of Federal Regulations, Part 51 Proposed Rule to Implement
The Fine Particle Ambient Air Quality Standards
•	EPA's National Ambient Air Monitoring Strategy
Our review of the speciation monitoring network did not include an independent
evaluation of the quality of the data generated by the monitors or an analysis of the
3
The Mid-Atlantic Regional Air Management Association is a voluntary, non-profit association of 10 State
and local air pollution control agencies that work together to prevent and reduce air pollution in the Mid-Atlantic
Region. Members include the States of Delaware, Maryland, New Jersey, North Carolina, Pennsylvania, Virginia,
and West Virginia; the District of Columbia; and Philadelphia and Allegheny Counties in Pennsylvania.
4
The Central Regional Air Planning Association is an organization of States, Tribes, Federal agencies, and
other interested parties that identifies regional haze and visibility issues and develops strategies to address them. The
Association, one of five Regional Planning Organizations across the United States, includes the States and tribal
areas of Arkansas, Iowa, Kansas, Louisiana, Minnesota, Missouri, Nebraska, Oklahoma, and Texas.
5	Formerly an acronym for "North American Research Strategy for Tropospheric Ozone," NARSTO is a
public/private partnership, whose membership spans government, the utilities, industry, and academia throughout
Mexico, the United States, and Canada. Its primary mission is to coordinate and enhance policy-relevant scientific
research.
4

-------
monitoring results. We evaluated the network's size, the location of the monitors,
the capabilities and limitations of the monitors, and EPA management of the
network. We did not evaluate emission inventories and atmospheric modeling; our
review of these components of the overall PM2 5 program was limited to how these
measurement tools are interdependent with the speciation monitoring network. We
conducted our evaluation in accordance with the Government Auditing Standards
issued by the Comptroller General of the United States. Our fieldwork was
conducted from March to October 2004. On February 3, 2005, we held a meeting
with officials from the Office of Air and Radiation (OAR) to discuss the Agency's
consolidated response to the draft report.
Prior OIG Coverage - Decline In EPA Particulate Matter Methods Development
Activities May Hamper Timely Achievement of Program Goals (Report No. 2003-
P-00016, issued September 30, 2003).
Results in Brief
EPA has made substantial progress in establishing a speciation monitoring network
that assists it in identifying and facilitating the development of control strategies
for sources of PM2 5, but still faces a number of challenges in ensuring that the
controls are implemented at the right sources. The development of control
strategies is best approached through collaborative processes that use emissions
inventories, ambient monitoring data, and air quality modeling. Although the
speciation network provides information on understanding the make-up and origin
of PM2 5, the Agency's ambient monitoring network does not fully assist in
providing the data needed for EPA or States to identify or quantify the chemical
make-up of PM2 5 particles, reliably trace particles back to their source, or account
for chemical changes that occur after particles are released into the atmosphere.
Speciation data are available to begin work on developing control strategies, and
EPA and the States are in the process of using the available monitoring data from
the Speciation, Supersites, and other State and private monitoring networks to
begin development of control strategies; however, increased monitoring efforts are
needed.
Rules promulgated under the Clean Air Act6 allow States that have PM2 5
nonattainment areas until February 2008 to develop control strategies for reducing
PM2 5, and an additional 2 years to reach attainment with the PM2 5 standard. Also,
with appropriate justification, the Act allows EPA to grant a State an extension of
up to 5 years, or until February 2015, to reach full attainment. Data from EPA's
speciation network will be vital to ensuring that pollution controls are implemented
6 Section 172 of the 1990 Clean Air Act required EPA to designate attainment based upon the National
Ambient Air Quality Standards. On July 1 8, 1997, EPA issued the rule, entitled "National Ambient Air Quality
Standards for Particulate Matter" (40 CFR Sec. 50.7), providing EPA with the authority to designate nonattainment
areas.
5

-------
at the right sources. Otherwise, some facilities may install unneeded controls,
while some needed controls may go uninstalled; ultimately, compliance may be
further delayed and more costly.
Agency officials acknowledge that improved speciation data will be needed for
EPA to overcome the uncertainties associated with PM2 5 particle origin. In 2004,
EPA budgeted over $43 million for PM2 5 monitoring, with about $16.4 million for
operation of the existing speciation monitoring network. However, only about
$800,000 was budgeted for improving the Agency's capability to address the
uncertainties with PM2 5 particle origin. According to manufacturers and some
Agency officials we contacted, increased partnering between EPA and monitor
manufacturers may be needed if advanced speciation monitors are to be developed
in time to help State and local agencies develop air pollution control strategies that
ensure controls are implemented at the right sources.
We recommend, among other things, that EPA increase its research on
technologies that can more fully assist in identifying the chemical make-up of
PM2 5 and, as such, account for the atmospheric impacts on PM2 5, and assay the
resultant changes that occur to the composition of the particle. This would include
greater attention to providing opportunities for cooperation with the private sector
to develop improved continuous speciation monitors. Detailed recommendations
are at the end of Chapter 3.
EPA's Office of Air and Radiation provided us with a response that consolidated
its comments to the draft report with those from the Office of Research and
Development (ORD). Although EPA disagreed with certain issues, the Agency
stated that the recommendations generally align with their current improvement
efforts. We included the Agency consolidated response in its entirety as
Appendix H. EPA did not agree with statements in the report that implied the
currently available speciation data was insufficient to help EPA and the States
"fully" develop effective control strategies. Nonetheless, our work with external
stakeholders, and key documents issued by NARSTO suggested limitations in the
available speciation data that would hinder EPA and the States from fully
developing effective control strategies to address the excess levels of PM2 5. Where
appropriate, we modified the report based on the Agency's consolidated response,
as well as several technical clarifications and comments also provided by EPA.
Our evaluation of the Agency's consolidated response is in Appendix I.
6

-------
Chapter 2
Speciation Monitoring Critical to Controlling
Fine Particulate Matter and Reaching Attainment
EPA and the States gathered 3 years of monitoring data to determine which areas
of the country are in nonattainment of the PM2 5 standard. The Federal Reference
Method (FRM) network consists of approximately 1,000 ambient air monitors used
to identify those areas of the country where people are exposed to unhealthy levels
of airborne PM2 5, and to what extent these areas exceed the PM2 5 standard. These
FRM monitors measure the mass, or weight, of the fine particles gathered on their
filters, but provide no information about a particle's chemical make-up. EPA uses
FRM data to determine whether areas are in nonattainment of the PM2 5 mass-based
standard. However, those areas in non-attainment will need, among other things,
speciation data to help them identify the source of the PM2 5 and determine the
chemical composition of the particle.
Nonattainment Areas Recommended
In February 2004, States and Tribes recommended to EPA those areas to be
designated as being in nonattainment of the PM2 5 standard. Nonattainment areas
are those with air quality levels exceeding the standards, plus nearby areas
contributing to such violations (also known as partial counties).
EPA evaluated the recommendations, and in some cases revised the State and tribal
submittals. In June 2004, EPA alerted 21 States that 244 counties, with a
collective population of 99 million people, were potentially nonattainment areas
for the new standard. This represents 35 percent of the Nation's total population of
285 million people. The nonattainment areas that EPA added to those initially
recommended by States and Tribes magnified the seriousness of the PM2 5 problem
nationwide, adding over 100 counties with 20 million people.
Table 2.1 provides details on the nonattainment areas recommended, and Chart 2.1
provides a map showing the location of those areas.
Table 2.1: Nonattainment Areas Recommended

Number of Counties
Total
Counties
Population Affected
By High Levels of
pm2.5
Full
Partial
States and Tribes
133
9
142
79 million people
EPA
233
11
244
99 million people
7

-------
Chart 2.1: Areas Anticipated To Be In Nonattainmenf
EPA Response to State Recommendations
on PM2 s Designations - June 29, 2004
a Note: this chart shows EPA and State recommendations for nonattainment areas based on their
analysis of the monitoring data; these are not the only areas where monitoring data exceeded the
PM2 5 standard one or more times in the last 3 years. Source: www.epa.gov/pmdesignations.
By February 2005, EPA is expected to officially designate those areas of the
United States that exceed the PM2 5 standard. These States have until February
2008 to develop control strategies for reducing PM2 5, and an additional 2 years to
reach attainment with the PM2 5 standard. Also, with appropriate justification, the
Act allows EPA to grant States an extension of up to 5 years, or until February
2015, to reach full attainment. As shown in Table 2.2, there are many activities
involved in developing and approving the State Implementation Plan (SIP) control
strategy, including public notice and comment, promulgating legally enforceable
State regulations, and in some cases enacting State legislation.
8

-------
Table 2.2: Key Activities and Milestones for Controlling PM25
Type of Activity
Milestone
EPA Designates PM2 5 Nonattainment Areas in Federal Register Notice
December 2004
Public comment period
December 2004 - February 2005
Nonattainment designations become final
February 2005
States Begin Developing SIP/Control Strategies
February 2005
EPA Issues Guidance to States on Development of Control Strategies3
June 2005
SIP Project Planning:
Define Scope, Develop Preliminary SIP Development Plan
Identify and Refine Key Program or Legal Issues and Resolve
Develop Technical Analysis and Inventory Preparation Plan
Data Gathering, Clarification, and Decisions
Finalize Preliminary SIP Development Plan
February 2005 - February 2008
SIP Development Phase:
Technical Data Gathering and Modeling
Control Strategy Development
Attainment Demonstration
Draft SIP Writing, Review SIP Draft, and Recommend Changes Required:
Implementation of All Reasonable Available Control Measures
Implementation of Reasonable Available Control Technology
Reasonable Further Progress
Comprehensive, Accurate, Current Emissions
Identification and Quantification of New Emissions Allowed
Permits for New and Modified Stationary Sources
Enforceable Emission Limitations, and Other Control Measures
Preparation of Contingency Measures
February 2005 - February 2008
SIP State and Local Adoption Phase:
Public Involvement and Formal Hearing Process
Finalize SIP, Adopt SIP, Submit to EPA
EPA Reviews SIP and SIP Development Schedule
February 2005 - February 2008
SIP Approval Phase:
EPA Receives SIPs
EPA Conducts Technical and Legal Review
Write Federal Register Notice
SIPs Approved as Final
February 2008 - February 2010
States Required to Reach Attainment
February 2010
Possible 5-Year Extension to Reach Attainment
February 2015
a Step out of sequence due to EPA delays in developing and issuing PM2 5 control strategy guidance.
While Table 2.2 depicts a number of key activities that EPA, State, local, and
tribal agencies will need to accomplish and the milestones they must meet in order
to reach attainment with the 1997 PM2 5 standard by 2010, it is not an all inclusive
list of activities. Appendix A shows the 23 major steps in the SIP development
and approval process.
9

-------
EPA has not yet set specific dates for when States must complete the many
interim activities to meet the February 2008 deadline for completing a control
strategy. After EPA makes designations, it is required to promulgate a schedule
defining when States must complete the interim steps necessary to meet Clean Air
Act requirements. Once nonattainment designations have been determined,
affected States must develop control strategies as part of their SIPs. A SIP
embodies the compilation of regulations, programs, and control strategies States
plan to implement to meet the NAAQS and reduce pollution to levels that meet
the health standard. Delays in reaching attainment goals may result in EPA
levying economic sanctions against States, such as withholding highway funds.
More importantly, delays in reaching attainment can result in tens of thousands of
premature deaths. Once attainment is reached, SIPs also help to maintain
acceptable levels of PM25.
From a public health standpoint, informed decisions made by EPA now and in the
next few years will be vital to ensuring that State, local, and tribal agencies meet
the milestones associated with reducing harmful levels of PM2 5. Although EPA
and State, local, and tribal agencies may have up to 10 years (2005 to 2015) to
lower the excess levels of PM2 5, many challenges must first be overcome.
Accurate and reliable speciation data will be necessary for EPA and State, local,
and tribal agencies to meet these Clean Air Act deadlines. Although the
speciation network provides some information for understanding the make-up and
origin of PM2 5, there are many other factors that also contribute to ultimately
reducing harmful levels of PM. EPA needs to ensure that it has: (1) a speciation
program that assists State, local, and tribal agencies to fully identify and quantify
the chemical make-up of PM2 5 particles, reliably trace particles back to their
source(s), and fully account for chemical changes that occur after particles are
released into the atmosphere; (2) robust and accurate PM2 5 emission inventories;
(3) effective atmospheric PM2 5 models; (4) timely and appropriate guidance
documents; and (5) timely and effective national controls.
Other Programs Impact PM2 5 Levels
In developing the control strategy through the SIP, EPA and the States also
consider emission reductions achieved from national programs directed at
reducing PM2 5. One of those national programs is the Acid Rain Cap-and-Trade
Program, which is designed to achieve emission reductions of Sulfur Dioxide
(S02) and Nitrogen Oxide (NOx), the primary causes of acid rain. A second
national program, the NOx SIP Call (formally known as Finding of Significant
Contribution and Rulemaking for Certain States in the Ozone Transport
Assessment Group for Purposes of Reducing Transport Ozone), requires 21 States
in the eastern half of the United States to revise their SIPs to help ensure that NOx
emission reductions are achieved to mitigate the regional transport of ozone
across State boundaries. Efforts to control S02 andNOx affect PM2 5 control
efforts because S02 and NOx can undergo a chemical reaction in the atmosphere
that transforms those pollutants into PM2 5. Also, there are two national programs
which are expected to reduce PM2 5 emissions from mobile sources - the Nonroad
10

-------
Diesel rule issued in June 2004 and the Clean Diesel Truck and Bus Rule
finalized in 2000. In addition, the Clean Air Interstate Rule, which is currently in
the proposed rulemaking phase, is expected to lower PM2 5 emissions. However,
EPA does not expect that national programs alone will bring nonattainment areas
into compliance with the PM2 5 standards. Details on these national control
programs are in Appendix B.
Speciation Data Critical to Overall Success of PM25 Program
To accurately identify the source of a PM2 5 particle, EPA and the States must first
understand its chemical make-up. Determining the chemical make-up of a
particle - known as "speciation" - is largely accomplished through data generated
by EPA's ambient air speciation monitoring program. Speciation data are an
integral part of the interdependent components that collectively comprise a
successful PM2 5 program. This is because the data allow EPA and States to
"groundtruth" other key aspects of the PM2 5 program, most notably their ability
to:
•	Gauge the accuracy and reliability of emission inventories.
•	Assess the validity7 of atmospheric, source-apportionment, and transport-and-
fate models and assumptions.
•	Measure the progress of national- and local-scale efforts to reduce PM2 5
emissions by providing data on the amount of sulfate, nitrate, ammonium,
organic and inorganic compounds, and other PM-related substances found in
the air.
EPA's Speciation Monitoring Program
As shown in Table 2.3, EPA's Speciation Monitoring Program consists primarily
of two networks:
•	The Speciation Trends Network (STN)8
•	The Interagency Monitoring of PROtected Visual Environments (IMPROVE)
Network
More details on these networks are in Appendix C.
7
We recognize that model validation is an iterative process that may never reach perfection; validation is
discussed here in terms of reducing model uncertainty and narrowing the range of possible outcomes, and in turn,
enhancing user and public confidence in model predictions.
g
The Speciation Trends Network (STN) consists of 54 trends monitors. There are also 215 State and Local
Air Monitoring Stations (SLAMS), not considered trends monitors, that provide speciated data. F or purposes of this
report, when we refer to the STN, we are referring to both the 54 trends monitors and the 215 SLAMS monitors.
11

-------
Table 2.3: PM25 Speciation Monitoring Program
Name
No. of
Monitors
Purpose of Network
Operated
By
STN
269
Generates data on chemical makeup of PM2 5 Capable of measuring
concentration levels of sulfate, nitrate, ammonium, and trace
elements including metals, elemental carbon, and organic carbon.
The STN is designed to complement the FRM network.
State and
Local
Agencies
IMPROVE
162
Measures visibility conditions, tracks visibility changes, and
determines causes for visibility impairment in U.S. National Parks and
Wilderness Areas. In 1999, IMPROVE network of about 50 monitors
increased to 162 monitors to supplement Regional Haze and PM2 5
programs. IMPROVE network monitors are mostly located in rural
areas, and provide measurements of regional and background levels
of PM2 5 concentrations. The same chemical components are
measured by IMPROVE as are measured by the STN, although
differences exist between the methods employed to collect and
analyze the collected sample.
National Park
Service and
other Federal
Land
Managers*
* Includes Forest Service, Fish and Wildlife Service, and Bureau of Land Management.
Several other networks are also used to monitor PM2 5 The previously noted FRM
network, consisting of 990 monitors nationwide, measures the mass of PM25, but
not its chemical make-up. The Photochemical Assessment Monitoring Station
(PAMS) network provides measurements of NOx, volatile organic compounds,
and ozone. The Clean Air Status and Trends Network (CASTNET) provides
measures of sulfate, nitrate, and ammonium ions, nitric acid (HN03), and S02.
Since the latter two measure NOx, S02, or HN03, all of which play a role in PM2 5
formation, some of the data generated from these two networks can also be
supportive of developing and measuring progress with PM2 5 emission reduction
strategies.
In Fiscal Year 2004, EPA budgeted approximately $43.75 million for PM ambient
air monitoring efforts. As shown in Chart 2.2, about $16.4 million was spent on
speciation monitoring between the STN and the IMPROVE network, but the
majority of those funds were dedicated to the operation of the current monitoring
network and the analysis of monitoring samples. Collectively, the funds budgeted
for improving both the existing speciation networks (STN and IMPROVE) totaled
about $800,000.
12

-------
Chart 2.2: FY 2004 Funding Budgeted for PM Air Monitoring (Total $43.75 million)
(5560,000 for Improvement ¦
represents 5% of
STN budget)
S12 m illion
S4.35 million
($217,500 for Improvem ent -
represents 5% of IMPROVE
budget)
S27.4 million
~ Other FM Monitoring ¦ STN ~ IMPROVE
To implement a successful PM2 5 attainment program, it is critical for EPA to have
speciation data that identify and quantify to the greatest extent possible the
chemical make-up of PM2 5 particles and associated uncertainties in the
measurements and, as such, assist in reliably tracing particles back to their
source(s), and fully assist in accounting for chemical changes that occur after
particles are released into the atmosphere. Knowing this information will better
enable EPA and the States to control the sources of PM25 and improve the quality
of the air. Also, ambient PM2 5 data are a primary method by which EPA and the
States measure progress toward attaining safe air quality. Below is a description
of how improved speciation monitoring data could improve EPA's PM2 5
emissions inventory and modeling efforts.
Speciation Data Used to Groundtruth Emissions Inventory Estimates
and Modeling Assumptions
EPA and the States primarily use three interdependent tools for managing its
PM2 5 programs: ambient air monitoring data, emissions inventory data, and
atmospheric modeling.9 These sources of information are used to make key
management decisions, prioritize issues, budget resources, measure progress in
meeting PM2 5 goals, and develop effective control strategies. Of the three
elements, monitoring data is the one most relied upon and recognized as the truest
depiction of what is occurring in the ambient air. As a result, unless the
improvement of monitoring data is a high priority to EPA, it will be limited in its
ability to help effectively control PM2 5.
9
Although we did not perform an evaluation of EPA's emissions inventory or atmospheric modeling, EPA
regional, State, and Regional Planning Organization (RPO) officials cited concerns about the reliability of these two
sources of information. Likewise, a review performed by NARSTO reached similar conclusions.
13

-------
Emission Inventories Verified With Speciation Monitoring Data
EPA maintains a database of air emissions information called the National
Emissions Inventory (NEI). NEI contains annual emission estimates of stationary,
area, and mobile sources of air pollution. With input from State, local, and tribal
air agencies, EPA amasses this inventory for PM2 5 and for each of the remaining
five criteria air pollutants,10 as well as estimates for the hazardous air pollutants,
also known as air toxics. EPA, NARSTO, and the States have described emission
inventories as the foundation to developing effective control strategies upon
which everything else is built. For details, see Appendix D.
There are a variety of methods used by industry and the States to develop these
emissions inventory estimates, some more reliable than others. One of the most
reliable is when a facility has a Continuous Emission Monitor (CEM) installed at
the point where the emissions are released into the atmosphere. These monitors
are designed to measure the emissions on a continual basis, thereby increasing
data accuracy and reliability. However, the majority of facilities do not have
CEMs installed and must therefore rely on less dependable methods of estimating
emissions. Often, facilities use emission factors generally developed by
monitoring emissions from several facilities and computing an average emission
rate, which is then applied to the entire industrial sector. This approach does not
always accurately account for differences in plant operations such as raw materials
used and equipment operated, and as a result is often recognized as unsuitable for
estimating an individual facility's emissions.
EPA regional, State, and RPO officials expressed concerns about the PM2 5
Emissions Inventory and how effective a tool it will be for developing control
strategies. For example, one State official said he had low confidence in the
existing inventories and that a good emissions inventory was, in his estimation,
two generations of inventories (at least 6 years) away from being sufficient to rely
on for developing control strategies. Likewise, previous studies performed by the
Government Accountability Office, EPA Office of Inspector General, and
NARSTO identified similar concerns.11
Through the use of air quality models, ambient monitoring data can be compared
to emissions inventory data to determine the level of consistency between data
sets and identify any discrepancies. When differences are identified, EPA
generally examines how the emission inventory can be improved, recognizing that
the monitoring data are the more reliable measure of atmospheric conditions.
10Six common air pollutants found nationwide that harm human health and the environment are called
criteria pollutants because EPA sets standards for these pollutants by first developing health-based criteria.
nGAO Report: EPA Should Improve Oversight of Emissions Reporting By Large Facilities (April 2001,
GAO-Ol-46). OIG Report: Decline In EPA Particulate Matter Methods Development Activities May Hamper
Timely Achievement of Program Goals (September 2003, Report No. 2003-P-00016). NARSTO Report: Particulate
Matter Science for Policy Makers - A NARSTO Assessment (February 2003).
14

-------
Emission estimates include not only PM2 5 but also data on some of the
components or precursors of PM2 5, such as volatile organic compounds or semi-
volatile organic compounds, S02, and NOx. Speciation data are used to determine
whether the emission estimates are consistent with what is being measured on the
filter. This comparison also helps EPA determine the extent to which PM2 5 may
have been secondarily formed in the atmosphere. The emissions inventory data
are also used as an input to atmospheric modeling. Therefore, the reliability of the
models - discussed in the next section - is dependent upon the accuracy of the
emissions data, which is verified in part by the speciation monitoring data.
Modeling Efforts Rely on Both Emissions and Speciation Data
Atmospheric models are the primary analytical tool in most air quality
assessments, which enables air quality managers to evaluate different emission
reduction scenarios to predict the impact of achieving desired reductions and
improving the air. The models allow model users to estimate the impact of
various factors influencing air quality, including meteorological conditions,
changes in emissions, and the effectiveness of proposed emission reduction
scenarios. They do this by making assumptions about the origin of particles, the
type of particles, the effectiveness of control strategies, and a host of other factors
called model inputs. For details on atmospheric modeling, see Appendix E.
There are two key models used for assessing the PM2 5 and precursor impacts of
various sources, also known as source apportionment, and these models are only
as accurate as the data (emissions and monitoring data) used to generate the
modeling results. Specifically:
•	Chemical Transport Models help EPA identify individual PM2 5 emitters at
the facility level by using data from the emission inventory, along with
meteorological conditions, to forecast changes in atmospheric concentrations
if emission rates change. These models are important because they provide a
means of linking primary PM2 5 and precursor emissions with those that are
formed secondarily in the atmosphere. Also, transport models assist in the
improvement of the emission inventories and ambient monitoring networks.
•	Source Receptor Models are not used to predict future ambient conditions, but
they explain events that have occurred, thus leading to an improved
understanding of the atmospheric impact. Receptor models start with inputs
of monitoring data, and work backward to determine the sources that are
contributing to levels of PM2 5 found at the monitoring sites, also known as
receptors. Receptor models help identify general source categories such as
diesel exhaust or coal-fired power plants, known as source apportionment.
EPA and State officials we contacted agreed that more analysis should be done in
the area of tracer species analyses, and that the source profiles found in the
SPECIATE database need to be updated. Further, NARSTO's February 2003
study concluded that similar improvements were needed. EPA recently initiated
15

-------
an update of the database that it plans to complete in 2005. EPA officials estimate
that this area could be improved within 3 to 5 years, if it receives the proper
funding and priority.
EPA, State, and RPO officials said there is room for improvement with regard to
the current models and, as the science progresses, the models should also
improve. NARSTO reported that chemical transport models can properly
represent the formation of sulfate. However, the formation of nitrate is more
difficult to represent because it requires knowledge of temperature, relative
humidity, and ammonia concentrations. The current state of scientific
understanding on the formation of secondary organic aerosols is insufficient, and
as a result PM2 5 modeling predictions at the present time have substantial
uncertainties. Continuous and semi-continuous speciation data would help
decrease these limitations.
Speciation Data Needed to Improve Understanding of PM Exposure
and Health Effects
Current NAAQS for PM are supported by findings from epidemiological studies
that have demonstrated positive associations between ambient PM mass
measurements and observed health impacts. As a result, the current PM NAAQS
uses particle mass as the indicator for the standard. However, there are questions
about the relative toxicity of various PM species and PM from various sources, as
well as whether a NAAQS that is based upon a metric other than mass is needed.
PM speciation data are needed to address these questions. Specifically, data are
needed to characterize the spatial (space) and temporal (time) patterns of PM
species and PM from various sources to improve our understanding of human
exposure to PM. This information in turn is important for epidemiological studies
that are investigating associations between observed health impacts and exposure
to PM species. However, the spatial and temporal gaps in the existing speciated
monitoring programs make it difficult to understand whether ambient
measurements of PM species are relatively homogeneous or heterogeneous across
space and time, which in turn makes it difficult to classify human exposure for
epidemiological studies. Enhanced speciated data are needed to address this
problem.
16

-------
Chapter 3
EPA Faces Many Challenges In
Identifying and Controlling Fine Particulate Matter
Although EPA has made substantial progress in establishing a speciation
monitoring network that assists it in identifying and controlling sources of PM25,
the Agency faces a number of challenges in implementing a program that ensures
that controls are implemented at the right sources. Although the speciation
network provides some information for understanding the make-up and origin of
PM2 5, the data provided by the network do not presently allow EPA and States to
identify and quantify the chemical make-up of PM2 5 particles to a sufficient
degree to reliably trace particles back to their source of origin, or fully account for
chemical changes that occur after particles are released into the atmosphere.
There are some speciation data available to begin developing control strategies,
but EPA and the States are not yet equipped with the speciation information
necessary to fully develop effective control strategies that take into account a full
understanding of PM25 chemical make-up, its sources, and the extent to which
each source contributes to overall PM2 5 levels.
EPA still has time to overcome these challenges, but increased efforts will be
needed to ensure that the States meet the milestones associated with reducing
harmful levels of PM2 5 expeditiously and at the least cost to industry. One of the
more promising approaches to obtaining information for better understanding,
tracking, and helping to control PM2 5 is the use of continuous and semi-
continuous monitors that would measure real-time PM2 5 levels. Semi-continuous
monitors for speciation are available for carbon, nitrate, and sulfate. However, to
date, continuous speciation data are limited, and improved speciation monitors are
needed to overcome this challenge. Increased partnering with PM2 5 monitor
manufacturers may help expedite not only the use of continuous monitor data, but
may also result in the development of other advanced speciation monitor methods.
Key Agency officials agreed that continuous and semi-continuous speciation
monitors would be the most likely approach to providing the robust data set
needed.
Limitations of Current Tools Used to Monitor and Assess
PM2 5 Air Quality
The speciation network, emission inventories, models, guidance documents, and
national control efforts are the primary tools EPA uses to operate its PM2 5
program. These tools provide EPA officials with the information for making key
management decisions, prioritizing issues, budgeting and allocating resources,
and measuring progress in meeting its PM2 5 program goals of reducing PM2 5 to
safe levels. Because EPA recognizes that these tools need improvement, the
17

-------
Agency is working to increase the reliability and usability of the information.
Table 3.1 elaborates on the uses and limitations of each of these tools.
Table 3.1: Uses and Limitations of Tools To Develop Control Strategies
Tool
Uses
Limitations
PM2 5 Speciation
Monitoring Program
Resulting data used to verify, evaluate,
improve, and "groundtmth" emission
inventories and models. Much is also
learned from data analysis efforts toward
understanding atmospheric processing and
accumulation of PM2 5 on local and regional
scales.
States need more and improved speciation monitoring
data to adequately identify and control sources of PM2 5.
Difficulties exist in adequately determining carbon
composition, such as the fractions and individual
species of carbon; accounting for secondary formation;
and obtaining real-time data. There are no standards to
assess uncertainty and biases in monitoring data.
Limited partnering with monitor manufacturers hampers
development of monitor advancements.
PM2 5 Emission
Inventories
Used to support monitoring activities and
as data input into atmospheric modeling.
EPA officials, States, and RPOs expressed concerns
about the reliability of emission inventories. For
example, the accuracy and representativeness often
are limited for emission factors and activity factors.
PM2 5 Models
Plans and predicts effectiveness of specific
control strategies and identifies relative
source contributions; helps to support and
improve monitoring data and emission
inventories; also assists in monitor siting.
EPA officials, States, and RPOs expressed concerns
about the reliability of models. Current modeling
predictions have large uncertainties. Much room for
improvement, especially as science progresses.
Guidance
Documents
Provides specific requirements, deadlines,
and suggestions to State, local, and tribal
agencies regarding development of control
strategies and related efforts to lower PM2 5
levels.
State officials must begin developing control strategies
in February 2005, but EPA does not plan to issue
guidance for developing the strategies until June 2005.
State officials anticipate the guidance will not be
specific enough to adequately assist the States.
National PM25
Controls
National efforts specifically designed to
address transport of PM2 5 across State
borders.
The Clean Air Interstate Rule and Nonroad Diesel Rule
will not be implemented until after the SIPs are due.
Ideally, all these tools should be fully developed today in order to place State,
local, and tribal agencies in the best position to begin developing their control
strategies in February 2005. State officials also expressed concern that they will
not have sufficient information to develop effective strategies within the time
allotted due partly to the limitations with each of these tools. As noted above,
EPA does not anticipate issuing the guidance for developing control strategies
until June 2005, eliciting concern among some State officials we contacted.
Challenges in Measuring Carbon and Ammonium, and Accounting
for Transport, Make it Difficult to Fully Identify and Quantify
Components of PM2 5
Generally, PM25 consists of six major components: sulfate, nitrate, ammonium,
organic carbon, elemental carbon, and crustal material, the latter estimated from a
group of trace metals that come primarily from soil. Because the composition of
PM2 5 differs in various parts of the country, it is important for States to know the
prevailing composition of PM25 in their area to adequately regulate the pollutant.
The speciation network is one of the most important tools available to the States
and EPA to estimate the composition of PM25. However, there are limitations
18

-------
associated with existing speciation monitors, and some officials are concerned
about how this may impact their efforts to ensure that controls are implemented at
the right sources. If not properly implemented, some facilities may install
unneeded controls, while some needed controls may go uninstalled. Ultimately,
compliance may be further delayed and more costly, and more people will suffer
adverse health effects longer.
An EPA regional official told us that in developing control strategies to address
local pollution problems, the State will need to manage with the data that exists.
The challenges presented to the States are greatly compounded by the fact that
over half of the PM2 5 is secondarily formed in the atmosphere. To fully identify
and quantify the components of PM2 5 and to accurately identify the source of the
particle, EPA and States will need to better understand:
•	The chemical make-up of the particle, especially the carbon components;
•	Chemical reactions that occur to the particle after it is released into the
atmosphere, especially ammonia's impact; and
•	The transport of PM2 5, wherein particles may travel considerable distances
from their point of origin depending on meteorological conditions and the size
of the particle.
EPA will need to increase its efforts in the research of PM2 5, including the
development of more sophisticated methods for monitoring and measuring PM2 5.
More Information Needed on Carbon Components to Accurately
Identify Sources of PM25
Carbon is arguably the most important constituent of PM2 5 and is estimated to
comprise between 30 and 60 percent of the total PM2 5 mass, with considerable
variability between locations and over time. EPA officials stated that because
carbon is a difficult pollutant to measure, sample, and analyze, not enough is
currently known about the carbon component of PM2 5. EPA officials said that as
PM2 5 levels decrease and as carbon becomes more prevalent, knowing more about
the carbon component will assist the Agency in evaluating emission reduction
strategies and gauging progress in reducing PM.
A senior Office of Air Quality Planning and Standards (OAQPS) official told us
that EPA knows enough about how much of PM2 5 is carbon, but they need more
information on what type of carbon compounds comprise the PM2 5 The official
explained that to better understand the carbon component of PM2 5, EPA needs to
know more about the two fractions of carbon - elemental and organic, as well as
the individual species that compose the organic carbon fraction. Accurately
measuring organic species is critical to help identify the source of up to 70 percent
of the particles, a key step in developing control strategies. For example, the
elemental carbon can be the result of forest fires or the combustion of diesel
19

-------
engines, while some organic particles originate from the combustion of gasoline
engines as well as from a number of natural sources (e.g., forest fires). Without
knowing the breakdown of the carbon component, EPA and the States' ability to
target the source of the pollution may be hindered. A senior OAQPS official said
there needs to be more work in organic chemistry if EPA is to better understand
the impact of the organic component on air pollution. He also said that this will
largely be accomplished through the development of new and advanced monitors
that measure the organic species of PM2 5. He did not believe the present monitors
employed by the STN and IMPROVE network would provide these data.
Knowing more about organic species in the air is important because it will further
EPA's understanding of atmospheric events as they relate to air pollution and
improve the Agency's capacity to model pollution events, evaluate emission
inventories, and better understand the link between sources and locations where
people are exposed to air pollution. In its February 2003 report, NARSTO also
emphasized the need for improved organic speciation monitoring, stating that:
Current organic speciation explains only 10 to 20 percent of total organic
compounds in the aerosol phase, and more work in this area is needed.
In agreement with NARSTO, the Agency noted that sufficient speciation of
organic aerosols will require improvements in both monitoring and analytical
capabilities. However, as noted above, the funds budgeted for improving the
existing STN and IMPROVE network were less than 5 percent of the ambient air
monitoring funds budgeted for the PM2 5 speciation program.
It should be noted that ORD is working with OAQPS, the research community,
States, the IMPROVE community and other Federal agencies, and academia to
better understand both collection and analysis methods for measuring carbon.
Examples of this collaboration include the Supersites Program and two recent
solicitations from the Science to Achieve Results (STAR) program, ORD's
extramural research grants program. Also, for the last 4 years, EPA has operated
an analytical laboratory dedicated to measuring organic compounds in PM. EPA
is also working collaboratively with the Supersites Program and the PM Health
Centers to better measure and characterize organic aerosols in an effort to identify
which compounds produce the highest risk to human health. However, our work
suggests that EPA will need to invest more in this complex area to sufficiently
understand the carbon component of PM2 5, where the particle originated, and the
potential impacts on human health.
Increased Effort Needed to Fully Understand the Impact of Ammonia
on PM2 5
PM2 5 is formed in two ways - primary formation and secondary formation.
Primary formation of PM2 5 occurs when a particle with a stable chemical form is
directly emitted into the air as a solid or liquid. This type of PM2 5 is more easily
traced to its original source because its components have not been altered since
20

-------
leaving the source. However, often over half of PM2 5 in the ambient air is a result
of secondary formation, which occurs when chemical reactions of gases in the
atmosphere either form new particles or condense onto other particles in the air.
Two common forms of secondarily formed PM2 5 occur when acid sulfates and
nitric acid react with ammonia in the atmosphere, creating ammonium sulfate and
ammonium nitrate, respectively.
EPA and State air quality managers told us that the sulfate-nitrate-ammonium
phenomenon presents a formidable challenge to them in identifying particle origin
and developing an effective control strategy. According to EPA studies, when
ammonium sulfate is collected on a Teflon filter, it typically absorbs moisture and
increases in mass. However, the Teflon filter is equilibrated under controlled
conditions to minimize the amount of water adsorbed to ammonium sulfate.
Overestimation of PM2 5 due to moisture adsorption is especially important
because the PM2 5 NAAQS standard is based on particle mass. The adsorption of
water for ammonium sulfate has been eliminated in the speciation monitoring
network due to the use of the nylon filter and analytical measurement techniques
used.
Conversely, the PM2 5 mass of ammonium nitrate may be underestimated because
nitrates are volatile in nature. EPA's Speciation Guidance document notes that
"nitrate losses during and after sampling have been well documented." Up to
50 percent of ammonium nitrate can be lost due to volatilization, generally due to
evaporation when the temperature rises during and after collection on the Teflon
filter. Studies are needed to determine how much nitrate is lost on the nylon filter.
Measuring ammonium nitrates is further complicated by the fact that once on the
filter, ammonium nitrate can break apart and return to its original compounds of
nitric acid and ammonia, depending on air temperature and humidity conditions.
The STN and the IMPROVE network address the volatilization loss with the use
of nylon filters which chemically bond with the nitric acid, thereby retaining the
deposit on the filter. However, ammonium is more complicated because the nylon
filter does not bond with the ammonia, which could result in volatility loss.
From an emissions inventory perspective, ammonia is the least understood of
these three interacting compounds. The largest sources of ammonia are generally
unregulated by EPA and the States at the present time and, as a result, EPA is
unable to determine the amount of ammonia emissions. Ammonia does provide
an environmental benefit in some instances by neutralizing acid (sulfuric and
nitric acids). Also, the costs and benefits of regulating ammonia are not fully
understood. For example, airborne ammonia is more likely chemically to
combine with acid sulfates over nitrate when more acid sulfate is available. When
sulfate is controlled and levels are reduced, the excess ammonia chemically
combines with nitrate instead to create ammonium nitrate. Although there is
some research being conducted in this area, unknowns regarding such ammonia
interactions impact EPA's ability to effectively control PM2 5 and make it difficult
to develop a clear strategy for reducing PM2 5 to safe levels.
21

-------
EPA officials told us that if EPA had more information on ammonia, the Agency
could better inform decisionmakers on the potential advantages and disadvantages
of regulating ammonia. As a result, EPA has taken steps to start an emission
inventory for ammonia. The speciation network could play a major role in
improving the Agency's understanding of ammonia by providing a constant
measure of this gas phase species through the speciation network. According to
an ORD official, measurements of ammonia and nitric acid, while desired, have
not been included in the network due to operational resources and cost. These are
gaseous, not particle, species, and therefore cannot be obtained from particle filter
measurements made by the Speciation network and require different sample
collection and analysis methods. However, the envisioned NCore Level 2 sites12
plan to measure ammonia and nitric acid as part of the multi-pollutant strategy.
More Accurate Accounting of Particle Transport Vitai to Determining
Source of PM25
One of the most challenging aspects of identifying and controlling PM2 5 is what
happens to the particle after it is emitted from the pollution source; whether it
comes from a stationary, area, or mobile source; or whether it is formed
secondarily in the atmosphere. Finding the source of particles is difficult not only
because the particles can change composition and molecular make-up, but
particles are also capable of regional transport. Particles may travel considerable
distances from their point of origin depending on meteorological conditions and
the size of the particle. Airborne PM2 5 has a lifetime of several days, enabling
particles to be carried hundreds and sometimes even thousands of miles in some
instances.
According to the NARSTO study, as well as EPA and State officials we
contacted, PM2 5 nonattainment problems typically result from a combination of
local source emissions and transported emissions from upwind areas. The Clean
Air Act requires that a SIP contain adequate provisions to prohibit sources in one
State from emitting air pollutants in amounts that contribute significantly to
nonattainment, or interfere with maintenance, in one or more downwind States.
To adequately address the transport issue and ultimately bring areas of the country
into attainment, EPA plans to combine simultaneous emission reduction efforts at
the local, regional, and national levels. However, with the annual cost of
compliance to industry estimated at more than $37 billion by the year 2010, there
are concerns that without improved speciation monitoring data on carbon,
ammonium/ammonia, and transport, some sources may dispute whether they are
the source of the PM2 5 emissions.
12
The NCore network is EPA's plan to repackage and enhance their existing ambient air monitoring
networks. EPA wants to more effectively leverage all of the existing major networks to produce an integrated
multiple pollutant approach to air monitoring. The overall structure of Ncore will range from the most complex
near-research grade sites (Level 1) to sites which measure only one pollutant (Level 3).
22

-------
EPA Has Made Efforts to Address Challenges in Obtaining
Speciation Data to Identify Sources and Develop Control Strategies
EPA has several efforts underway to enhance its PM2 5 speciation monitoring
capabilities, including: (1) attempts to supplement the STN data with speciation
data from the IMPROVE Network; (2) a continuous speciation monitoring pilot
study to assess the viability of these continuous monitors supplementing the
speciation data obtained from the STN; and (3) various speciation monitoring
research efforts to better understand the unknown characteristics of PM2 5, such as
the PM Supersites Program. The Agency's work in each of these areas is critical
for EPA and the States to effectively manage their PM2 5 programs. To make the
progress necessary to control excess levels of PM2 5 within the time frames
mandated by the Clean Air Act, EPA will need to more vigorously pursue existing
projects and undertake new efforts to improve its ability to effectively regulate
PM2 5. Below is a description of EPA's ongoing efforts, the benefits derived, and
the work still needed to adequately identify sources of PM2 5 and facilitate the
development of effective control strategies to reduce PM2 5 to safe levels.
Supplementing STN With IMPROVE Data Provides Some Useful
Information But Compatibility Is Limited
One way EPA is trying to obtain sufficient speciation data to identify sources of
PM2 5 is by taking advantage of data generated from the IMPROVE network, a
collection of 162 rural monitors operated primarily by the Department of Interior's
Federal land management agencies, along with State, local, and tribal agencies.
However, because the purpose, design, and desired results of the IMPROVE and
STN vary significantly, there are mixed views within EPA regarding the extent to
which IMPROVE can supplement the STN, and, as such, differing views on the
level of effort that should be expended in trying to make the two networks
compatible. Table 3.2 shows key differences in how the monitors were
manufactured, where they are located, and how they are operated.
23

-------
Table 3.2: Comparison of IMPROVE and STN Networks
Key Features
IMPROVE
STN
Location
Rural Areas
Urban Areas
Sampler Design
One design for the entire
network
Samplers are provided by four
manufacturers, each with its own
design, but they have been shown to
be comparable for most species.
Approximately 90 percent of the
monitors were supplied by one of the
four manufacturers.
Operation &
Maintenance
Trained federal land managers
Professional air quality monitoring
technicians
Frequency of Filter
Replacement
Every third day
Every third day
Shipping & Handling
Regular mail - no temperature
control. ORD and OAQPS are
involved in a study to determine
if the shipping method impacts
the amount of semi-volatile
material collected on the filters.
Preliminary results are expected
soon.
Shipped by a commercial carrier -
cold storage to preserve accuracy of
results.
Measurement
Differences
Utilizes specific blank correction
techniques for reporting carbon
data.
Does not utilize specific blank
correction techniques for reporting
carbon data.
Started in 1985 with about 50 monitors, the IMPROVE network was designed to
help monitor visibility in the U.S. National Parks; as such, the IMPROVE
monitors were sited in rural areas throughout the United States. The STN was
established in 1999 by EPA regulation as a companion to the mass-based FRM
network, which measures particle mass for compliance with the PM2 5 NAAQS
health-based standard but does not speciate particles. Because EPA's principal
focus is protection of public health, the STN is located in highly populated urban
areas, contrary to the rural IMPROVE monitors.
While the design is different between the two networks (rural vs. urban), both
networks collect and measure the same species (see Table 3.2). However, there
are differences in the collection and analysis methods which can result in
differences in the reported concentrations by each type of monitor. In addition,
the IMPROVE network measures light-absorption and light-scatting because of
their importance to understanding visibility degradation in clean areas. Filters
from both the STN and IMPROVE monitors are collected and analyzed every
3 days; however, officials told us that continuous data are also important because
it provides real-time data.
EPA and some RPOs are conducting comparison studies to identify whether the
data from the two networks are comparable. In 2001, EPA selected six locations
(three urban and three rural) and sited an IMPROVE monitor next to an STN
monitor to determine the compatibility of the two networks by comparing how
24

-------
closely the monitoring results matched. At the time of our work, there was not a
firm completion date for the comparison study; however, the ORD official
responsible for overseeing the study told us that the first 2 years of data will be
presented to EPA and State officials at a conference scheduled for February 2005.
During 2004, EPA phased in the implementation of an additional 9 urban sites,
and expects data to become available beginning in the Fall of 2005.
Although preliminary findings from the study have shown similar results for
sulfate measurements, EPA and State officials have concerns about the
compatibility of the two networks primarily because of the differences in some of
the PM2 5 constituents each network measures. For example, EPA has found that
because the STN and IMPROVE monitors use different methods for measuring
carbon, the amount of elemental carbon is overestimated by the IMPROVE
method and underestimated by the STN method. Likewise, there are some
differences in how sulfate is reported. However, a senior OAQPS official said
that based on the first year of results from the six-city study, the impact of these
differences is negligible.
The difficulty of finding a method to make the carbon data of the two networks
agree has fueled the discussion of what should be done in the interim to help EPA
and State, local, and tribal agencies identify sources and develop effective control
strategies. Views are mixed within EPA on how to proceed with integrating the
IMPROVE network and the STN. One viewpoint is that the most practical and
logical approach would be to move toward more consistency and compatibility by
replacing the STN with new IMPROVE samplers. In 1999, State officials and
others within EPA believed that because the IMPROVE monitor was designed for
use in rural areas, where pollution levels are generally lower, it was still uncertain
how the monitors will perform in urban areas. However, EPA officials stated that,
over time, this became less of a concern because initial results showed the
IMPROVE monitors functioned well in urban areas. Still, there are some EPA
officials who believe that instead of replacing the STN, it is more reasonable to
continue their ongoing efforts to make the data of these two networks more
compatible.
EPA is working to better understand the uncertainties associated with the two
methods, including ORD's program for examining the analytical differences
between the STN method and the IMPROVE method for measuring carbon. To
clearly understand the extent to which these two networks are compatible, EPA
will need to invest more in identifying the differences between the two networks
and resolving the uncertainties identified. Answers to these questions will be
important in designing and implementing effective pollution control strategies.
Continuous Speciation Monitor Pilot Study May Help Identify Sources and
Develop Control Strategies
Monitor manufacturers and State and EPA officials we contacted said that
continuous speciation monitors are among the more promising near-term
25

-------
technologies for understanding the components of PM25. In July 2002, EPA
began a pilot study to test three types of continuous monitors - nitrate, sulfate, and
carbon - to determine how compatible these continuous monitors were with the
STN monitors, and to what extent the monitors can be used to supplement the
STN. The study was scheduled to be completed by July 2005 and initially planned
for 12 sites, with each site having the three continuous monitors sited beside an
existing STN monitor. In August 2004, the OAQPS official responsible for
performing the study told us that EPA had deployed continuous monitors at 5 of
the 12 sites, is planning to add 3 sites by the end of 2004, and then add the
remaining 4 sites in 2005.
EPA acknowledged that the study is behind schedule, and explained that there
were some modifications needed on the monitoring equipment after the monitors
were deployed. The OAQPS official also said that monitor manufacturers have
been reluctant to address some of the equipment malfunctions because it is
uncertain whether EPA will commit to ordering more of the continuous monitors
in the future. The results of the study will help States develop effective control
strategies; however, delays in completing the study may adversely impact the
development of these strategies. As noted above, States must begin developing
control strategies in February 2005, 4 months before the pilot study is scheduled
to be completed. At that time, it appears that EPA will only have partial results on
how continuous monitors can supplement the STN.
As shown in Table 3.3, for the three types of monitors (carbon, sulfate, and
nitrate), the findings from the pilot study have been mixed.
Table 3.3: Status of EPA Pilot Study of Three Continuous Monitors
Monitor Type
Advantages / Limitations
Status
Continuous Carbon
Provides real-time data. Measures lower than the STN
samplers at all sites.
July 2005s
Continuous Sulfate
Provides real-time data. Compares well with the STN
samplers when the concentrations of sulfate levels are
low. Does not compare as well at higher sulfate levels.
July 2005s
Continuous Nitrate
Provides real-time data. Consistently lower than the
STN sampler measurements; however, there is better
agreement at lower nitrate levels.
July 2005s
a As noted above, EPA is behind schedule with the study due to problems encountered
deploying the monitors and the monitor manufacturers making the necessary modifications.
By the end of the 3 years (July 2005), it is not likely that EPA will have collected the amount of
data originally anticipated.
In addition to continuing the pilot study, EPA officials said they plan to work with
monitor manufacturers to seek improvements of the existing monitors and to
evaluate different continuous monitors at new sites. Through discussions with
EPA and the monitor manufacturers, we found that there is room for improved
communication among both parties. For example, the Chief Executive Officer of
one leading monitor manufacturer told us that it would be helpful to them if EPA
26

-------
were to more clearly articulate its plans for speciation monitors, so that
manufacturers could prioritize their own research and development efforts. With
improved communication, both parties could better resolve problems identified
and more quickly implement the continuous speciation strategy. EPA officials
said that when acceptable performance has been demonstrated with continuous
monitors, the Agency will increase the use of real-time data on the amount of
sulfate, nitrate, and carbon found in ambient air. This will help identify particle
origin and develop effective control strategies.
Supersites Studies Will Increase Understanding of PM2 5
EPA has several ongoing efforts to better understand and characterize the
properties of PM2 5, with the PM Supersites Program being by far the most
prominent effort. The PM Supersites Program is a $26.5 million ambient
monitoring research program designed to compare and evaluate different
monitoring methods, as well as testing new and emerging measurement methods
that may ultimately advance the scientific community's ability to measure the
physical and chemical components of PM in the air. The program is intended to
help address the scientific uncertainties associated with the measurement of
ambient concentrations and atmospheric processes, and source-receptor
relationships of PM2 5. The Supersites Program also was designed to support
health and exposure studies, but not fund them directly. EPA established eight
Supersites in 1999 and 2000 through 5-year cooperative agreements with leading
atmospheric sciences universities throughout the United States. The PM
Supersites Program results should help EPA and the States with their efforts to
develop effective control strategies.
With regard to understanding continuous monitors, the Supersites Program is
evaluating several different continuous monitors by performing side-by-side
comparisons at several of the sites. The studies include not only identifying
differences in the results of the monitors, but also trying to understand why these
differences occurred. For more details on the Supersites Program, see
Appendix F.
ORD recognizes that the Supersites Program findings will aid EPA and States
with their efforts to develop effective control strategies. As shown in Table 3.4
below, two of ORD's Annual Performance Measures (APM) under the
Government Performance and Results Act (GPRA) relate to EPA's efforts in
identifying and controlling PM2 5.
27

-------
Table 3.4: Key EPA Research and Development GPRA Measures
ORD Annual Performance
Measure Under GPRA
Description
ORD 2005 APM"
Deliver to OAR, States, and the scientific community data from the
Supersites program via an internet accessible large relational
database that can be utilized for air quality model evaluation and to
perform integrated analyses of data from the various Supersites
locations.
ORD 2006 APM"
Deliver to OAR and the States results from the Supersites program
that can be used to prepare and evaluate SIPs [including control
strategies]
a Completion of the APM is expected no later than September 30, 2005.
b Completion of the APM is expected no later than September 30, 2006.
ORD's PM Research Program Includes Efforts to Improve Speclatlon
Data and Source Identification
ORD is conducting research to address some of the limitations and challenges
related to measuring and modeling speciated PM and improving source
apportionment. Resources to address these issues in ORD's in-house research
program total approximately S3 million annually and include:
•	Organic aerosol sampling and analysis methods development;
•	Development and application of receptor modeling tools;
•	Analysis of PM Elemental composition by X-Ray Fluorescence;
•	Development of chemistry modules for secondary organic aerosols and aerosol
nitrates;
•	Improved measurement methods for elemental carbon/organic carbon and for
speciating organic PM;
•	Evaluation of continuous methods for sulfate, nitrate, and carbon and related
precursor species;
•	The Detroit Exposure and Aerosol Research Study - includes research to
understand human exposure to PM species and PM sources and to understand
spatial and temporal distributions of PM species;
•	PM Supersites Program;
•	Development and evaluation of the Community Multiscale Air Quality chemical
transport model, including specific improvements in predicting organic PM and
nitrates; and
•	Inverse modeling to improve ammonia emission inventories.
ORD's extramural grants program, Science to Achieve Results (STAR), recently
issued two related grant solicitations. The first solicitation, Measurement,
Modeling, and Analysis Methods for Airborne Carbonaceous Fine Particulate
Matter, resulted in 16 grant awards totaling approximately $6.6 million. The goal
of this grant solicitation is to conduct research that would improve measurement
methods, models, and analysis techniques used to quantify emissions and ambient
28

-------
concentrations of PM25. The focus is on studies that will provide insights in, and
improved techniques to quantify, the organic and elemental carbon fractions of
PM2 5 and to more fully understand the specific chemical species that make up the
organic fraction. The research results from the solicitation will allow
identification of significant sources of PM2 5 and enable EPA, State, local, and
tribal agencies to design more effective and efficient air quality management
plans. According to ORD officials, early results, such as journal articles, are
expected by late 2005 and final results by 2008. The titles and institutions for
each of these grants are included in Appendix G.
The second STAR solicitation, Source Apportionment of Particulate Matter, is
expected to result in 11 grant awards totaling approximately $4.5 million dollars.
The award of these grants is anticipated by early 2005.
Increased Partnering with Monitor Manufacturers Could Improve
Monitoring Capabilities and Uses
EPA and State officials agree that there is a need for increased use of continuous
speciation monitors that provide real-time information on the PM2 5 collected at
the monitoring site. Real-time data provide a more accurate depiction of what is
occurring in the atmosphere, accounts more fully for meteorological impacts, and
better pinpoints the sources of PM2 5. Likewise, continuous monitoring data will
provide the scientific community and the regulatory community with more useful
data in overcoming the challenges associated with understanding and controlling
PM2 5. However, EPA and State officials, as well as monitor manufacturers,
acknowledge that continuous speciation monitors need improvement before these
monitors can effectively provide what is needed. Although EPA is currently
conducting a pilot study to test continuous monitors in coordination with a major
monitor manufacturer, both EPA and monitor manufacturer officials we contacted
agreed that increased effort in developing effective continuous monitors is needed
to meet the needs of the State, local, and tribal users.
EPA awards limited funding to monitor manufacturers for the development of
new and improved monitors. However, an overall lack of partnering between
EPA and PM2 5 monitor manufacturers has contributed to limited progress in
developing and deploying continuous monitors that would measure real-time
PM2 5 levels. The limited partnering is attributed, in part, to the manufacturers
not having sufficient assurance that the research and development efforts will be
worth the investment. OAQPS officials said that this lack of assurance results
from the low number of monitors that EPA is often requesting. For example,
EPA is interested in obtaining monitors for 12 sites, with several monitors at each
site, totaling 30 to 40 monitors. OAQPS officials said that this volume of
monitors does not provide EPA much leverage in influencing the manufacturer's
use of research and development resources. Monitor manufacturers we contacted
told us they are seeking improved communications and more commitment from
29

-------
EPA through both increased funding provided by the Agency and more certainty
that there will be sufficient demand for the monitor when it is ready for use.
The Chief Executive Officer of one leading ambient air monitor manufacturer told
us there was a need for stronger EPA leadership in developing partnerships with
monitoring manufacturers, which would help his company prioritize the most
worthwhile and promising projects to develop. He told us that, with a limited
budget, his company must select between EPA and other governmental agencies
as to what they believe to be the most worthwhile research and development
projects for them to pursue. With regard to development of an improved monitor,
the Chief Executive Officer said that EPA is becoming a lower priority because
the Agency is not doing enough to involve and lead the monitoring manufacturers
to the research and development areas that would be most promising. For
example, the manufacturer has considered discontinuing the research and
development of two promising continuous and near continuous speciation
monitors because of a lack of partnering with EPA. One effort involved the
development of a monitor that would record hourly measurements of specific
metals frequently found in PM2 5 The second effort was to develop a continuous
monitor to measure ultrafine particles.
A senior OAQPS official agreed that increased partnering was an "excellent
suggestion," and that by providing manufacturers with increased research and
development funding, the Agency would demonstrate a stronger commitment to
the effort. Importantly, he noted that in his view EPA would need the more robust
data provided from continuous monitors if EPA and the States are going to fully
understand the chemical composition of particles, understand what happens to a
particle after it is released in the atmosphere, and trace a particle to its origin.
These are the steps necessary to ensure that States develop effective control
strategies.
Conclusions
In September 2004, EPA researchers not only reconfirmed the previously
identified serious health effects of exposure to excess levels of PM2 5, but also
found that such exposures adversely impact the heart. EPA officials agree that
this is an acute health problem that needs to be addressed expeditiously. The
Agency has made substantial progress in establishing a speciation monitoring
network to aid in identifying and developing controls for sources of PM2 5, but still
faces a number of challenges in implementing a program that ensures that the
controls are implemented at the right sources. With estimates of annual control
costs to industry exceeding $37 billion by 2010, EPA is likely to face substantial
implementation challenges unless the Agency can to a greater extent identify and
quantify the chemical make-up of PM2 5 particles and, as such, reliably trace
particles back to their source of origin, and account for the changes that occur
after particles are released into the atmosphere. Despite several years of effort,
30

-------
EPA's STN does not presently do this with sufficient certainty for all the
constituents of PM25.
EPA still has time to overcome these challenges, but increased efforts appear to be
needed to ensure that these data will be available to help State, local, and tribal
agencies meet the milestones associated with reducing harmful levels of PM2 5
expeditiously and at the least cost to industry. It is highly important that controls
be implemented at the right sources. Otherwise, some facilities may install
unneeded controls, while some needed controls may go uninstalled; ultimately,
compliance may be further delayed and more costly.
Recommendations
Due to the adverse health effects occurring annually from PM2 5, as well as the
estimated S3 7 billion annual compliance cost in the year 2010 to industry, we
recommend that the Assistant Administrator for Air and Radiation, in
collaboration with the Assistant Administrator for Research and Development,
expand and expedite EPA's efforts to overcome the challenges associated with
identifying and controlling PM2 5 In particular, we recommend that EPA:
3-1 Increase from 5 to 10 percent the OAQPS funding allocated for
performing analytical assessments, adopting new methods, and conducting
research on technologies that can more fully assist in identifying the
chemical make-up of PM2 5, account for the atmospheric impacts on PM2 5,
and assay the resultant changes that occur to the composition of the
particle, with particular emphasis on:
a)	Increasing and improving the speciated data for the six major
components of PM2 5 (sulfate, nitrate, ammonium, organic carbon,
elemental carbon, and crustal material). This could be accomplished
largely by the increased development and use of continuous speciation
monitors.
b)	Enabling EPA and State, local, and tribal agencies to perform more
sophisticated analyses, through source-receptor modeling and other
analysis and modeling methods, to better identify the source of the
PM2 5 and fill the gaps in the data generated from the STN and
IMPROVE networks.
3-2 Identify the uncertainties associated with the comparability of similar
speciation monitoring methods, such as the IMPROVE and STN methods,
and develop short- and long-term plans to address these uncertainties and
increase the usability of the data generated from the various speciation
networks. Specifically:
31

-------
a)	Complete the six-site comparability study and incorporate the results
of the study into Agency decisionmaking.
b)	Expedite Agency efforts to determine whether the STN and IMPROVE
monitors can produce adequately comparable data and, if not,
determine which method should be further deployed to increase data
consistency.
3-3 Increase Agency efforts to develop the data needed to conduct the more
advanced analyses necessary to understand the behavior, characteristics,
and chemical composition of PM2 5, including:
a)	Increasing efforts to develop methods to collect and measure source
profiles at emissions sources, and the respective tracers in ambient air
that uniquely identify those sources.
b)	Identifying and minimizing the uncertainties associated with
measuring the organic fraction of PM2 5.
c)	Adding the capability to measure ammonia to the ambient Speciation
Network.
d)	Developing and deploying continuous speciation monitors that help
provide the real-time data needed to more accurately depict what is
occurring in the atmosphere on a real-time basis and better pinpoint the
sources of PM2 5.
3-4 Address the challenges described in Recommendations 3-1, 3-2, and 3-3
by establishing a new workgroup or through an existing workgroup,
comprised of officials from OAQPS, ORD, and selected EPA regions;
State, local, and tribal agencies; State and Territorial Air Pollution
Program Administrators/Association of Local Air Pollution Control
Officials; RPOs; affected industries; academia; and monitor
manufacturers.
3-5 Through the workgroup discussed in Recommendation 3-4, increase
partnering efforts with monitor manufacturers to maximize the availability
and use of current continuous speciation monitors and expedite the
development of the next generation of speciation monitors to address the
challenges described above. Given the health and economic consequences
if controls are not implemented expeditiously and at the right sources,
EPA should consider a joint EPA-private sector pre-competitive
technological research program similar to the groundbreaking Partnership
for a New Generation of Vehicles (PNGV) program that helped to develop
a new generation of low emitting vehicles.
32

-------
Agency Comments and OIG Evaluation
EPA made detailed comments on our draft report and, where appropriate, we
made revisions. The Agency generally agreed with the recommendations in the
report, but disagreed with the statements in the report that referred to the
speciation monitoring network's inability to help EPA and the States to fully
address the PM problems. With respect to recommendation 3-1, the Agency did
not agree that EPA and the States could not develop control strategies. We agreed
that some control strategies could be developed by EPA and the States, but some
challenges still exist. We continue to believe that improved data from EPA's
speciation network will be vital to ensuring that pollution controls are
implemented at the right sources. Otherwise, some facilities may install unneeded
controls; some needed controls may go uninstalled; and, ultimately, compliance
may be further delayed and more costly. The Agency's consolidated response and
our evaluation of that response are in Appendices H and I, respectively.
33

-------
Appendix A
Major Steps in SIP Development and Approval Process
a-
if
Source: "State Implementation Plan Process Improvement Project Final Report: Recommendations for improving
the development and approval of State Implementation Plan (SIP) revisions in EPA Region 10," April 15, 2002.
34

-------
Proposed SIP DEVEU)PMMM AND APPROVAL PROCESS
Nun - A t tainroent/V isibilitv SIP (cominn.d i

xirmynwmtsi phase
: -ja"	f
II
12
TECHVL M DU

OVsJPOL
1

SIP WAFT
GvT~W\tl K

>11' \TF'i rN
, hF-VH ^T^KA

WRITING
\1 (HiHiVi
>
D^vtior-lfM
1





1--EPA
STATE AM) LOCAL
ADOPTION PHASE
L-S/L
ir
t! Bl |C
LI V LVLN r
. EORM U HFARfMi
'RU* f Si
r SUBMITTAL
TEMPLATE
S', Vit fi ILKM
UTOPIIO*
*tP
si-BMirm.
HAL MCE
HKLPAKAIION
L^L
C»EPA
n'\
l*s/l
L=.VL-
REVlfcWSJP
DRAW
RECOMMEND
changes
FINALIZE
srp
DISPLTE
RE-SOLUTION j
D'-rr MFNT
LEARNM
CHECK
UEPA
SIP APPROVAL PHASE
Ij=EPA
Ui^f IE
CCWLFTWSS
HP H lvT
l' vO,
U - [!b il I K,K
TO sTATT.S
gUV IKw AK D
S ?fi ^T*-" -W
RL\ 1LM
sr^rmn£
TH HMr M aMj
I tC M -^LV tLW

COMFLEn.
-THltr
A=HPA
L-LFA
L=tn
* >/L
L-LfA
L—VL
risiinrAfi^N
ranrsiLAL
L₯X M, POLmOi
i
WRITE RtDCftAL
REGISTER
KOTtCE
St \ t fc
LLTIbK
hi I 'JrMbN J
Plan
ltpa
lv=EPA
TR H. SU'POfc
UUC-MLN
JLA2NIM
fWr'K
REGISTHF
NOT1CT

-------
Appendix B
National Emission Control Programs Providing Benefit to
PM2 5 Emission Reduction Efforts
EPA has several ongoing and planned national efforts specifically designed to address the
transport of PM2 5 across State borders. State, local, and tribal agencies will need to consider
how, if at all, these national efforts will impact local PM2 5 emissions and how these national
initiatives will affect the development of control strategies. A description of five of the more
significant national efforts likely to result in benefits to the PM2 5 program follows.
NOx SIP Call
The rule, Finding of Significant Contribution and Rulemaking for Certain States in the Ozone
Transport Assessment Group Region for Purposes of Reducing Regional Transport of Ozone,
commonly known as the NOx SIP Call, was issued final on October 27, 1998. This regulation
required 22 jurisdictions (21 States and the District of Columbia) in the eastern half of the United
States to revise their SIPs to help ensure that NOx emission reductions are achieved to mitigate
the regional transport of ozone precursors across State boundaries. The NOx SIP Call
rulemaking requires that these 22 jurisdictions adopt and submit SIP revisions that contain
provisions adequate to prohibit sources from emitting NOx in amounts that can contribute
significantly to nonattainment of the 1-hour and 8-hour ozone national ambient air quality
standards. The States were required to tighten controls on NOx-emitting facilities during the
ozone season, which covers a 5-month period from May 1 through September 30 each year.
States had to be in compliance by May 31, 2004, with the exception of Georgia and Missouri,
which must comply by May 1, 2005. The coal-fired electric utility industry, believed to be a
significant contributor to PM2 5 emissions, was one of the industrial sectors that was required to
comply with the NOx SIP Call.
Acid Rain Cap-and-Trade Program
As part of a two-phased approach, the Clean Air Act set a goal of reducing annual S02 emissions
by 10 million tons below 1980 levels by placing a cap on the total emissions from a select group
of the nation's largest fossil fuel-fired power plants. Phase I began in 1995 and affected 263
units at 110 mostly coal-burning electric utility plants located in 21 eastern and midwestern
States. An additional 182 units joined the first phase of the program as substitution or
compensating units. Emissions data for 1995 indicate that S02 emissions at these units were
reduced nearly 40 percent below their required level. Phase II, which began November 2000,
tightened the annual emission limits imposed on these large, higher-emitting plants, and also set
restrictions on smaller, cleaner plants fired by coal, oil, and gas, encompassing over 2,000 units
in all. The acid rain cap-and-trade program affects existing utility units serving generators with
an output capacity greater than 25 megawatts, as well as all new utility units.
The acid rain cap-and-trade program introduced an allowance trading system that uses the
incentives of the free market to reduce pollution. Under this system, affected utility units are
36

-------
allocated allowances13 based on their historic fuel consumption and a specific emissions rate.
Allowances may be bought, sold, or banked. Anyone may acquire allowances and participate in
the trading system. This second phase set a permanent ceiling of 8.95 million allowances for
total annual allowance allocations to utilities.
Clean Diesel Truck and Bus Rule
In 2000, EPA finalized the Clean Diesel Truck and Bus Rule. This rulemaking provides for the
cleanest-running heavy-duty trucks and buses in history. Due to new control technologies and
cleaner, low-sulfur fuel, these vehicles are expected to be over 90 percent cleaner than today's
trucks and buses. Engine manufacturers will have flexibility to meet the new standards through a
phase-in approach between 2007 and 2010, with new low sulfur diesel fuel provisions expected
to go into effect by June 2006. When fully implemented, diesel soot emissions will be reduced
by nearly 110,000 tons each year, resulting in preventing 8,300 premature deaths, preventing
1.5 million lost work days, 7,100 hospital admissions, and 2,400 emergency room visits for
asthma every year, according to EPA.
Clean Air Nonroad Diesel Rule
Nonroad diesel emissions currently account for about 47 percent of total diesel PM and about
25 percent of total nitrogen oxides of the combined on-road and nonroad diesel emissions
nationwide. The Nonroad Diesel rule issued in June 2004 is expected to reduce sulfur in fuels
for diesel engines by 99 percent by setting tighter emission standards for diesel engines used in
construction, agricultural, and industrial equipment. The rule sets standards for new engines that
will be phased in beginning in 2008 with the smallest engines, and moving on to larger ones,
until all but the very largest diesel engines meet both NOx and PM standards in 2014. Some of
the largest engines (greater than or equal to 750 horsepower), will have one additional year to
meet the emissions standards. When the full inventory of older nonroad engines has been
replaced, the nonroad diesel program will annually prevent up to 12,000 premature deaths, one
million lost work days, 15,000 heart attacks, and 6,000 children's asthma-related emergency
room visits, according to EPA.
Planned Clean Air Interstate Rule
Similar to the acid rain and NOx SIP call, the proposed Rule is designed to be abroad cap-and-
trade approach to reducing emissions. The proposed rule targets S02 and NOx emissions from
power plants, which significantly contribute to pollution problems in other downwind States.
These pollutants lead to formation of PM2 5 and ground-level ozone that are associated with
thousands of premature deaths and illnesses each year. EPA asserts that full implementation of
this proposal would reduce S02 and NOx emissions by approximately 70 percent from pre-
implementation levels. EPA plans to propose this Rule by December 2004.
13
Each allowance permits a unit to emit 1 ton of S02 during or after a specified year.
37

-------
Appendix C
Description of the Two Primary
Ambient Speciation Networks
The STN and the IMPROVE network are the main components of the speciation program and
generate the same type of speciation data using similar sampling and analytical approaches,
collecting the samples on filters which are then analyzed for speciation results. State, local and
tribal agencies deploy and operate the STN, while the Federal land managing agencies of the
Department of the Interior operate the IMPROVE network.
Speciation Trends Network (STN) - The STN was initiated in 1999 to generate data on the
chemical make-up of fine particles and to determine trends in concentration levels of selected
ions, metals, carbon species, and organic compounds in PM2 5. Mostly sited in urban areas, the
STN monitors were placed at 269 locations across the nation, of which 54 trends monitors were
sited by EPA and 215 were sited by State, local, and tribal agencies. Four manufacturers produce
almost all of the STN monitors. Each manufacturer's speciation monitor varies in design and the
method in which PM2 5 is collected. The STN is intended to complement the activities of the
much larger PM2 5 FRM mass-only monitoring network used for attainment designations and aid
in the development of control strategies to lower PM2 5 levels to within the NAAQS standard.
Some of the other uses of the STN are to:
•	Develop annual and seasonal spatial characterization of aerosols;
•	Conduct air trends analysis;
•	Track the progress of control programs; and
•	Integrate the STN data with the IMPROVE data.
IMPROVE (Interagency Monitoring of PROtected Visual Environments) - Maintained and
operated by the National Park Service, the IMPROVE network was initiated in 1985 as a
long-term monitoring program to measure visibility conditions, track changes in visibility, and
determine the causes for visibility impairment in the U.S. National Parks and wilderness areas.
Prior to 1999, the IMPROVE network consisted of about 50 monitors across the country, but in
1999 was expanded to supplement the Regional Haze and PM2 5 programs. The IMPROVE
network largely consists of 162 monitors located in rural areas providing measurements of
background levels of PM2 5 emissions. EPA is exploring ways that the IMPROVE network can
supplement the STN to provide additional speciation information in support of developing
control strategies. However, EPA has not yet determined the extent to which the IMPROVE
monitors can augment the STN.
38

-------
Appendix D
PM2 5 Emissions Inventory
and Its Relationship to Monitoring and Modeling
Once every 3 years EPA prepares a national database of air emissions information with input
from the State, local, and tribal air agencies, and from industry. Known as the National
Emissions Inventory (NEI), the database contains an estimate of the annual emissions from
stationary and mobile sources that emit criteria air pollutants (PM2 5 being one of these) and their
precursors, as well as hazardous air pollutants. The NEI includes emissions information for all
50 States, estimated at the individual point level for major sources (facilities), as well as county
level estimates for area, mobile, and other sources. The NEI database sorts emissions into three
source categories:
•	Major - stationary sources of emissions, such as an electric power plant, that can be
identified by name and location.
•	Area - small point sources, such as a home or office building, or a diffuse stationary
source, such as a wildfire or agricultural tilling.
•	Mobile - any kind of vehicle or equipment with a gasoline or diesel engine, including
cars, trucks, airplanes, and ships.
The emission inventory provides an estimate of source data, some of which are speciated, for
what is emitted into the atmosphere, while the speciation network provides speciated data of
what is in the atmosphere. The two data elements are compared to identify major discrepancies
and to determine how the emission inventory can be improved. Also, by identifying major
differences in PM2 5 levels when comparing emission inventory data and monitoring data, EPA
learns more about the extent of secondary formation of particles.
Emission inventories are a good place to start to develop control strategies. They have been
described as the foundation, upon which everything else is built. The source speciated data
generated by the emission inventory are input into the chemical transport model to help predict
the effects of various control strategy scenarios. The receptor models use speciated monitoring
data as an input to help verify the accuracy of emission inventories. However, except for data
from Continuous Emissions Monitoring Systems, emission inventories are largely estimates of
sources' emissions based on emission factors and activity or usage profiles, whereas speciation
monitors measure actual emissions found in the ambient air.
39

-------
Appendix E
Atmospheric Modeling
and Its Relationship to Monitoring and Emissions Inventory
Atmospheric models have become a primary analytical tool in most air quality assessments and a
critical component in developing effective control strategies. Different air pollutants are closely
related because they experience similar atmospheric processes and can often originate from the
same source. For example, knowing more about how automobiles affect ozone, acid rain, and
PM2 5 formation will lead to a better understanding of how changes in the level of one pollutant,
or its precursors, may lead to the changes and concentrations of another. Models enable air
quality managers to run different emission reduction scenarios to predict what impact they would
have on the air pollutants of concern, if their modeling assumptions are correct.
Models for assessing the impacts of PM25 basically fall into two categories: chemical transport
models and source receptor models. As shown in Table E-l, these two models use different
approaches to help provide decisionmakers with the information needed to decide what emission
controls are needed to lower PM2 5 levels. It is the use of both types of models that help
decisionmakers understand the impacts of both primary (best done by receptor models) and
secondary (best done by chemical transport models) aerosols.
Table E-1: Uses of Source Receptor and Chemical Transport Models
Receptor Models
Chemical Transport Models
Receptor Oriented (Monitoring Sites)
Source Oriented
Identifies Sources
Identifies Sources
Estimating Contributions From Those
Sources
Predict Changes In Future Concentrations
Evaluating Emission Inventories
Evaluating Emission Inventories
Helps Plan The Application Of
Chemical Transport Models
Helps Select Sites For Monitoring
Helps Evaluate And Improve
Chemical Transport Models Results
Helps Link Secondary Aerosols To Sources
Increases Understanding Of
Atmospheric Environment

Application To Secondary Aerosols
Limited

Diagnostic Model

Receptor Models
Ambient air monitoring data are input into receptor models to help trace the particles found on
the monitor back to the sources that are emitting the PM25. As a result, the reliability of the
40

-------
model is directly dependent on the quality of the ambient air monitoring data. When used in
conjunction with information on emissions, receptor models can identify areas of the emission
inventories that need improvement. The receptor models are not used to predict future ambient
conditions, but instead help to explain events that have occurred. However, the receptor models
cannot fully account for the chemical processes occurring in the atmosphere that result in
secondarily formed PM2 5. Receptor models alone cannot characterize and quantify the
relationships between the sources and the monitoring data. The best available approach currently
for addressing the uncertainties in source-receptor relationships involves the application of
multiple techniques. For example, use of receptor models, along with emission inventory data,
speciation monitoring data, and chemical transport models will lead to a better understanding of
PM2 5 sources and formation.
Two key analyses performed to target and develop emission control strategies are source
attribution and source apportionment, which both involve the identification of possible sources
and the level of contribution. Source attribution identifies individual PM2 5 emitters such as a
specific refinery, while source apportionment identifies general source categories such as diesel
exhaust and chemical manufacturers.
Chemical Transport Models
Chemical transport models are important tools because they help to link primary and precursors
of secondary PM2 5 emissions to ambient concentrations, and can assist in the improvement of the
emission inventory and the location of ambient monitors. These source-oriented models use data
from the emission inventory along with meteorological conditions to forecast changes in
atmospheric concentrations if emission rates change. Specific uses of chemical transport models
include:
•	Primarily used to predict changes in ambient concentrations in emissions control
scenarios, and thus, to allow more effective and efficient means for reducing PM
levels through emission reduction strategies.
•	Estimating the contributions of local and long-range sources.
•	Finding potential errors in emission inventories (when modeled results differ
significantly from ambient PM2 5 concentrations, further analysis is required to
determine whether the emission inventory inputs are more likely the cause of the
discrepancy than the model).
•	Assisting in the design and evaluation of PM2 5 monitoring networks (because of the
model's ability to make independent estimates of current atmospheric conditions,
designers of PM2 5 networks can use this information to evaluate how well the
deployed network is representing the predicted results).
Before chemical transport models can fulfill their valuable roles, they must be evaluated using
ambient speciation data and under various conditions to demonstrate their ability to predict
atmospheric conditions. However, chemical transport models can never exactly represent
atmospheric conditions, because of limitations in the scientific understanding of atmospheric
processes.
41

-------
Two other analyses performed by EPA that help the Agency better understand the speciation of
PM are described below.
Source Profile and Tracer Species
Two important types of data fed into these models are source profiles and tracer species. Using
speciation monitoring and emissions data, source profiles are in-depth analyses conducted at one
plant or facility where the various chemicals being emitted are speciated. Tracer species analyses
involve finding inert elements that are emitted from a plant or facility that do not change
composition in the atmosphere. As a result, EPA is able to trace the known particle from the
monitor back to the source. For example, some forms of coal contain unique elements allowing
experts to determine the type of coal burned, enabling them to pinpoint sources by tracing which
facilities burn that type of coal. Tracers allow EPA to have more confidence in their modeling
results.
Back Trajectory Analysis
An important secondary tool or method used in conjunction with the source apportionment
models is back trajectory analysis because it helps to improve and refine the source
apportionment modeling results. Back trajectory analysis is a technique that incorporates data on
sources contributing to the PM2 5 levels measured by a receptor, with meteorological data (i.e., air
flow patterns) to determine the likely source location. Continuous speciation monitors would
impact this tool positively by assisting in comparing source and meteorological data to speciated
ambient data.
42

-------
Appendix F
Advanced PM2 5 Research Conducted by
Supersites Program
Each Supersites Project uses a mixture of routine and advanced measurement methods to better
characterize the chemical and physical properties of PM in the air. The methods include
measurement of PM mass and the chemical components of PM in a variety of size ranges,
physical properties, precursor species, and related variables such as meteorological parameters.
The methods development focuses on better understanding how and why PM accumulates in the
air and relating PM at a receptor site back to its sources.
EPA officials describe the Supersites program as an in-depth characterization of PM in those
Regions with the highest PM concentrations. Several of the specific studies being conducted at
Supersites Projects directly address some of the challenges EPA faces in PM speciation. For
example, at several sites, researchers are studying atmospheric measurements to characterize PM
constituents, atmospheric transport, and source categories that affect the PM in their region,
although some aspects might be generalized to other regions. This information is essential for
understanding source-receptor relationships and the factors that affect PM at a given site (e.g.,
meteorology, sources, transport distances). This information is also essential for improving the
scientific foundation for atmospheric models that investigate PM accumulation, exposure, and
risk management questions. Other closely related efforts include:
•	Comparing and evaluating different methods of characterizing PM (e.g., emerging
sampling methods, routine monitoring techniques, and the FRM);
•	Testing new and emerging measurement methods that may ultimately advance the
scientific community's ability to measure the physical and chemical components of
PM in the air and, thereby, better understand the process affecting PM in the air and
to investigate exposure and health effects;
•	Quantifying the impact of the various sources (transportation, power plants, etc.) on
the PM concentrations in the area, including those locally generated versus those
transported from upwind areas, which may be regional in nature; and
•	Supporting regulatory agencies in the development of emissions reduction
implementation plans that cost-effectively reduce particle concentrations on urban and
regional scales.
ORD recognizes that the Supersites Program findings will largely benefit EPA and the States'
with their efforts to develop effective control strategies. As part of reporting on its GPRA goals
and measures, ORD has set two APMs that directly relate to EPA's efforts in identifying and
controlling PM2 5:
ORD 2005 APM - Deliver to OAR, States, and the scientific community data from
the Supersites program via an internet accessible large relational database that
43

-------
can be utilized for air quality model evaluation and to perform integrated
analyses of data from the various Supersites locations.
ORD 2006 APM - Deliver to OAR and the States results from the Supersites
program that can be used to prepare and evaluate SIPs [including control
strategies]
Two other APMs have been associated with the Supersites Program. Both focus on synthesizing
findings associated with the development and evaluation of continuous monitors for mass and
chemical components of PM. One of these, APM 25 was completed in 2003 and the other is a
2006 deliverable. EPA is planning another key outreach meeting in February 2005, entitled
"2005 AAAR [American Association for Aerosol Research] PM Supersites Program and Related
Studies." This is an international specialty conference with a focus on results from the Supersites
Programs and other methods, measurements, modeling, and data analysis studies conducted
during the last 5-7 years.
44

-------
Appendix G
Recent Grant Awards from ORD's STAR Solicitation
Grant Title
Institution
Evaluation and Minimization of Organic Aerosol Sampling Artifacts Using
Impactors and Quartz Fiber Filter Denuders
UC- Riverside
Application of Thermal Desorption GC-MS for the Analysis of Polar and Non-Polar
Semi-Volatile and Particle-Phase Molecular Markers
UW-Madison
Advancing ATOFMS to a Quantitative Tool for Source Apportionment
UC- San Diego
Integrating the Thermal Behavior and Optical Properties of Carbonaceous
Particles: Theory, Laboratory Studies, and Application to field Data
Univ of Illinois
Atmospheric Processing of Organic Particulate Matter: Formation, Properties,
Long Range Transport, and Removal
Carnegie Mellon
Secondary and Regional Contributions to Organic PM: A Mechanistic
Investigation of Organic PM in the Eastern and Southern United States
Rutgers
Source-Oriented Chemical Transport Model for Primary and Secondary Organic
Aerosol
UC Davis
Development of Advanced Factor Analysis Methods for Carbonaceous PM
Source Identification and Apportionment
Clarkson
Secondary Aerosol Formation from Gas and Particle Phase Reactions of
Aromatic Hydrocarbons
UNC-Chapel Hill
Fundamental Experimental and Modeling Studies of Secondary Organic Aerosol
Caltech
Emissions Inventory and Process Reconciliation Using Molecular Markers and
Hybrid/Inverse Photochemical Modeling with Direct Sensitivity Analysis
Georgia Tech
Understanding Thermal and Optical Carbon Analysis Methods
DRI
Particle Sampler for On-Line Chemical and Physical Characterization of
Particulate Organics
MIT
Atmospheric Aerosols from Biogenic Hydrocarbon Oxidation
Univ of Colorado
Development and Application of A Mass Spectra-Volatility Database of
Combustion and Secondary Organic Aerosol Sources for the Aerodyne Aerosol
Mass Spectrometer
UC-Riverside
Aetahlometric Liquid Chromatographic Mass Spectrometric Instrument for
Characterization of Carbonacceous Ambient Particulate Matter. Laboratory and
Field Studies
Texas Tech

45

-------
Appendix H
Consolidated EPA Response to Draft Report
o
%
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
JAN 3 1 2005
MEMORANDUM
OFFICE OF
AIR AND RADIATION
SUBJECT: Response to the Draft Evaluation Report: EPA Needs to Direct More Attention,
Efforts, and Funding to Enhance Its Speciation Monitoring Program for
Measuring Fine Particulate Matter, Assignment No. 2003-1450
FROM: Jeffrey R. Holmstead
Assistant Administrator
TO:
J. Rick Beusse
Director for Program Evaluation, Air Quality Issues
Thank you for providing us the opportunity to respond to the draft report from the Office
of Inspector General (OIG) issued December 3, 2004. The purpose of this memorandum is to
provide comments on the draft evaluation report, "EPA Needs to Direct More Attention, Efforts,
and Funding to Enhance Its Speciation Monitoring Program for Measuring Fine Particulate
Matter, Assignment No. 2003-1450." This response has been coordinated with EPA's Office of
Research and Development (ORD).
The recommendations provided by the OIG generally align with our current
improvement efforts. Our concerns with the OIG report pertain to: 1) characterization of
the current state of affairs, and 2) the need to balance resources across all aspects of the air
program. We disagree with negative statements in the report regarding the sufficiency of
currently available speciation data to "fully" develop effective control strategies.
Nevertheless, EPA recognizes that improvements are clearly needed in our current inventory,
monitoring, and modeling programs to further improve the efficiency and credibility of control
strategies.
OAR is experiencing a reduction in budget and expects to see limited funding in the
coming years. With anticipation of static staff resources, the competing needs on our other
monitoring networks, Biowatch, and the implementation of the National Ambient Air Monitoring
46
See
Appendix I
Notes 1 and 2

-------
Strategy (NAAMS), we will need to prioritize efforts put forth on these recommendations.
Prioritization will allow OAR to focus on those recommendations deemed most critical. We will
consider the OIG final recommendations along with expected recommendations from the Clean
Air Act Advisory Committee Air Quality Management review, and related recommendations
received on an ongoing basis from Clean Air Scientific Advisory Committee's subcommittee on
ambient air monitoring and methods.
General comments are provided in the attached response, along with several specific
comments that are more technical in nature. ORD will include their comments as a marked-up
copy of the report that will be provided separately to OAR and the OIG via email.
If you have additional questions or require clarification, please contact Peter Tsirigotis of
my staff at (919) 541 -9411.
Attachment
cc: Pete Cosier, Office of Air and Radiation, Audit Follow-up Coordinator (6102A)
Dr. Dan Costa, National Human and Environmental Effects Laboratory (B143-02)
Thomas C. Curran, Deputy Director, Office of Air Quality Planning and Standards
(C404-04)
Dr. Gary J. Foley, Director, National Exposure Research Laboratory (MD-75)
Lek G. Kadeli, Acting Deputy Assistant Administrator for Management (8101R)
Ardra Morgan-Kelly, Audit Liaison, National Exposure Research Laboratory (D343-01)
William Lamason, Associate Director, Emissions, Monitoring and Analysis Division
(C304-02)
Phil Lorang, Leader, Ambient Air Monitoring Group (D243-02)
Stephen D. Page, Director, Office of Air Quality Planning and Standards (C404-04)
Joann Rice, Ambient Air Monitoring Group (D243-02)
Dr. Rich Scheffe, Emissions, Monitoring and Analysis Division (C304-02)
Dr. Linda Sheldon, Director, Human Exposure and Atmospheric Sciences Division
(E205-01)
Dr. Paul Solomon, Human Exposure and Atmospheric Sciences Division (D205-03)
Laurie Trinca, Audit Liaison, Office of Air Quality Planning and Standards (C404-2)
Peter Tsirigotis, Director, Emissions, Monitoring and Analysis Division (C304-02)
James Vickery, National Exposure Research Laboratory (D305-1)
Timothy Watkins, Deputy Director, Human Exposure and Atmospheric Sciences Division
(E205-01)
47

-------
Attachment
We have the following general comments on the draft conclusions and recommendations:
General Comments
(1) We object to the statement in the draft report that "EPA and States are not yet
equipped with the necessary information to fully develop effective control strategies. "
The following statement is made or implied throughout the draft report (e.g., on pages 5,
13, and 17, and in the "At a Glance" section):
"Although some speciation data is available to begin work on developing control strategies, EPA
and the States are not yet equipped with the necessary information to fully develop effective
control strategies."
This position is not supported by the findings of others in the air quality science and policy
communities. For example, the NARSTO community, which includes all major U.S. Federal,
State and private sponsors of air quality research, in its 2003 report, "Particle Matter Science for
Policy Makers - a NARSTO Assessment, " points out that "Policy makers are currently benefiting
from research initiated five to ten years ago, or longer. This research provides a basic
understanding on PM formation, transport, and its major contributing sources. It characterizes
the areas of North America where PM concentrations, visibility reduction, and potential
population exposure are the greatest. Despite considerable uncertainties, sufficient scientific
confidence exists to devise management actions likely to improve air quality (emphasis added)."
Corresponding comments have been received from the Clean Air Scientific Advisory Committee
(CASAC) Ambient Air Monitoring and Methods Subcommittee in their review of the National
Monitoring Strategy and have been among Clean Air Act Advisory Committee's (CAAAC)
recommendations on improving air quality management.
The OIG report correctly points out (p. 13) that "EPA and States primarily use three tools
for managing its PM2 5 programs: ambient monitoring data, emissions inventory data, and
atmospheric modeling." However, EPA does not use these tools independently. The
report's conclusion and affirmation it attributes to NARSTO that "unless improvement of
monitoring data is a high priority to EPA, it will be limited in its ability to help effectively
control PM2 5" are seriously overstated and in error. As the NARSTO assessment points out
(Synthesis, p. 24), "Source specific options to reduce PM concentrations are best approached
through corroborative analysis using emissions inventories, ambient concentration measurements
and air quality modeling. Given the strengths and limitations of any one of these science tools, it
is recommended that they be used in an integrated manner..." No one of these three tools is more
important than another. As an example, EPA integrated emissions inventories, modeling and
speciation monitoring data when it evaluated the impact of regional S02 and NOx as part of the
Clean Air Interstate Rule (CAIR) [www.epa.gov/air/interstateairquality], OAR and ORD are
working together to address these issues through a variety of research efforts supported by PM
Supersites, STAR Grants and other extramural activities, some of which are outlined in the OIG
report on pages 44-45.
See
Appendix I
Note 1
See
Appendix I
Note 2
48

-------
The OIG report makes the case that determining the chemical make up of PM is largely
accomplished through data generated by EPA's ambient air speciation program, and the
draft lists EPA's two principle networks, the STN and IMPROVE, with total investments
of S16.5M yearly. Only brief mention is made of the 5-year $26.5 M Supersites program
instituted to apply state-of-the-art monitoring and speciation methods to particle
characterization. The report omits the fact that the Supersites program is an in-depth
characterization of PM in those regions of the U.S. with the highest PM concentrations. EPA has
established the program to address the very thing that the OIG has recommended. In fact, the
Supersites program is the subject of a specialty conference by the American Association for
Aerosol Research next month. This conference and the subsequent policy-relevant
recommendations, through ORD's 2006 Annual Performance Measure (APM), will be very
helpful to the Agency in making potential improvements to the monitoring program. Further, the
OIG report should recognize the equally large measurement studies sponsored by States and
private industry in California (CRAPACS) and the Southeast U.S. (SEARCH), and by other
Federal agencies such as NOAA, DOE, and NASA in the Northeast U.S. (NEAQS). These
studies collectively have produced a wealth of PM speciation information that will well equip
States with the necessary information they need to develop effective control strategies.
(2) The draft report inadequately describes the role of speciation monitoring for
developing "effective" control strategies and the work being done at EPA.
The roles of the Speciation program are to provide data for:
•	assessing the effectiveness of emission reductions strategies through the
characterization of air quality trends;
•	supporting the development of predictive modeling tools and the application of source
apportionment modeling for control strategy development;
•	supporting programs aimed at improving environmental welfare, such as the Regional
Haze program; and
•	supporting health effects and exposure research studies.
This information is valuable in crafting control strategies to address the principal sources of PM
problems, as well as to assist in better understanding the components of PM that are of greatest
significance to human health effects.
The statement in item (1) also refers to developing "effective" control strategies.
However, it is very difficult to say what is, and what is not, an "effective" control strategy.
The real question is what pollutants do we need to reduce to minimize risk from PM?
From that information, we need to develop effective control strategies. EPA has, in fact,
implemented controls and reduced PM levels considerably. Lead from gasoline has been
eliminated. Sulfate in the East has dropped due to S02 controls and nitrate also will likely drop
due to the NOx SIP call, as well as the Acid Rain Program. Further reductions may come from
the proposed CAIR. We can identify the major sources (power plants, cars, etc) and address a
big part of the PM problem, but once again the question is, are they the right sources to reduce
the risk from PM? This leads to the need for speciation data to improve our understanding of the
relative toxicity (and resulting risks) from various PM sources. In our response to the OIG
position papers, we suggested adding a section/paragraph entitled "Speciation Data Needed to
Improve Understanding of PM Exposure and Health Effects," which has been incorporated on
page 16 of the draft report. The point of this suggested paragraph was to highlight the fact that to
See
Appendix I
Note 3
See
Appendix I
Note 4
See
Appendix I
Note 5
49

-------
develop more "effective" control strategies, we need to understand what characteristic of PM
drives the observed health impacts. In others words, is it particle size, composition, species or
some combination that leads to health impacts? Speciated data are needed to support exposure
and health research to answer these questions in addition to developing control strategies for
PM2 5, which is the emphasis of the report. The bottom line is that the most effective control
strategy will consider the sources of PM that are responsible for the greatest health risk in
addition to reducing PM2 5 mass to meet the PM National Ambient Air Quality Standards
(NAAQS).
(3) The report does not address the need to balance research priorities within the air research
program and across other media.
The principal recommendation of the OIG report is that EPA needs to direct more
resources to speciation monitoring. As pointed out in item (1) above, speciation
monitoring is only one of three tools needed for PM NAAQS implementation; the other
two are emissions characterization and air quality process understanding and modeling.
EPA has carefully balanced its investment across all three tools to address the key
remaining uncertainties. The need to fully integrate and balance these three tools is a poii
in the National Resource Council's (NRCs) concluding report on PM research priorities (
IV). Revisiting of this balance through our annual allocation process may be warranted, 1
major shift that would come at the expense of the other two areas of research would be
inappropriate and would not serve the interests of enhanced air quality management. It is
important to note that EPA must balance research investments supporting these three area
(monitoring, emissions characterization, and air quality modeling) with research needs in
areas of exposure, health effects, and control technology development.
As part of the Agency's annual planning and budgeting process, ORD works with EPA's
other program and regional offices to allocate funds across various research programs.
This process ensures that media-specific recommendations are fully considered and that
the areas of greatest need are given the highest priority. Using this process, the OAR has
an opportunity to elevate the relative priority of research supporting PM speciation
monitoring. It is important to note, however, that ORD is already making significant investments
in this area of research with results and research products anticipated in the near future. Finally,
ORD must balance EPA's needs for research not only within the air research program, but also
across all environmental activities.
Responses to the Recommendations
3-1 Increase from 5 tolO percent the OAQPS funding allocated for performing analytical
assessments, adopting new methods, and conducting research on technologies that can more
fully identify the chemical make-up of PM25, account for the atmospheric impacts on PM2S,
and assay the resultant changes that occur to the composition of the particle, with particular
emphasis on:
OAR supports the general intent of the recommendations. However, we are not endorsing
the specific recommendation regarding the funding increase, which does not account for
competing priorities in the air program. It is important to note that ORD also allocates
funding to conduct research to address these issues.
See
Appendix I
Note 6
nt made
Report
but a
»also
as
. the
See
Appendix I
Note 6
See
Appendix I
Note 6
50

-------
a)	Increasing and improving the speciated data for the six major components of PM2S
(sulfate, nitrate, ammonia, organic carbon, elemental carbon, and crustal material).
To the extent that the recommendation implies equal attention to improvements for all six
components across the nation, we disagree. This recommendation lacks specific detail
regarding what is meant by "increasing and improving" the speciation data. Please clarify
what is meant by "increasing and improving the speciated data." For example, does this
refer to the number of sites (i.e., collect more spatial data) or their geographic distribution;
does it refer to higher time resolution (i.e., implementation of continuous methods); or does it
refer to measuring a larger number of species (i.e., focus on the organic species and methods of
analysis with better limits of detection for the inorganic species)? It would be an inefficient and
unproductive use of scarce resources, for instance, to increase speciation sampling for pollutants
in parts of the nation where reliable emissions information indicates there are few or no
significant sources. Also, we can measure all the species specified within a certain set of
uncertainties, so is the report asking for improved methods that will reduce the uncertainties in
the measurement methods? We recognize that there is room for improvement in some of the
speciation methods currently used, especially for carbon measurements. ORD currently has
several efforts underway to address these issues.
b)	Enabling EPA and State, local, and tribal agencies to perform more sophisticated analyses,
through source-receptor modeling, to better identify the source of the PM2mS and fill the
gaps in the data generated from the SIN and IMPROVE networks.
We believe that the tools (emissions, modeling, and measurements) currently available to
identify the sources of PM 2 5 are sufficient for developing effective control strategies for
attainment of the NAAQS. Given the measurements that are available today, the current
receptor modeling tools are capable of providing a broad characterization of the sources
contributing to ambient PM2 5 levels which can be used for developing effective control
strategies. One potential complication is the level of expertise available, particularly in the State,
local, and tribal agencies, to apply these tools. As a result, any additional near term investments
may better be directed at developing and delivering guidance for applying source apportionment
techniques, particularly receptor modeling approaches.
While we believe that current receptor modeling tools are capable of supporting control strategy
development; improvements in our measurements and modeling tools will certainly improve our
ability to more specifically identify sources of PM. For example, to be able to separate and
identify additional specific sources, detailed measurements (e.g., hourly measurements conducted
on a daily basis as opposed to 24-hour integrated averages conducted on a l-in-3 day basis) and
improved modeling tools to take advantage of these measurements, would be needed. However,
these enhancements would require substantial additional investments, well beyond the 5 to 10
percent suggested in this recommendation. EPA is committed to advancing the science in this
area and has a program to develop improved source-receptor tools, but as stated previously,
investments in this area must be balanced with investments in other priority research areas.
As noted in the OIG report (pp. 15-16), EPA is investing in improvements to emission source
profiles by updating of the speciation source profile database (SPECIATE), planned for
completion in 2005. SPECIATE will be an important resource in source apportionment studies.
See
Appendix I
Note 7
See
Appendix I
Note 8
51

-------
3-2 Identify the uncertainties associated with the comparability of similar speciation
monitoring methods, such as the IMPROVE and STN methods, and develop short- and long-
term plans to address these uncertainties and increase the usability of the data generated from
the various speciation networks. Specifically:
a)	Complete the six-site comparability study and incorporate the results of the study into
Agency decision making.
Besides the initial 6-site study, there have been an additional nine STN/IMPROVE sites
added to assess comparability and informing network decisions. This information will
also be used to develop a plan for future collocated sites to help understand the differences
between the data generated. ORD plans to present data analysis results at the upcoming
American Association for Aerosol Research meeting in Atlanta, Georgia, in February
2005. OAR is beginning the task of compiling the results from the first 6-site study and laying
out questions specifically directed at informing the decision making and program improvements.
b)Expedite	Agency efforts to determine whether the STN and IMPROVE monitors can
produce adequately comparable data, and if not, determine which method should be further
deployed to increase data consistency.
See comment above. In addition, ORD has research underway that is targeted at identifying the
"optimal" thermal-optical analysis method as noted in Appendices F and G of the OIG's report.
Results from that research can aid in the identification of the method best suited for future
deployment.
3-3 Increase Agency efforts to develop the data needed to conduct the more advanced
analyses necessary to understand the behavior, characteristics, and chemical composition of
PM2 5, including:
a) Increasing analytical work related to source profiling and tracer species, such as
fingerprinting carbon to its original source.
Please clarify whether this recommendation addresses emissions-related monitoring,
ambient-related monitoring, or both. An emissions-related recommendation would
address "source profiling," while an ambient monitoring recommendation would address
the measurement of "tracer species" in air, as opposed to the source. Since both source
profiling and tracer species are mentioned, it could be assumed that the recommendation
addresses both emissions and ambient monitoring. However, it is important to understand what
(or how) source profiles are used and the relationship between source profiles and tracer species.
Tracer species are unique markers for a source which are identified by measuring source profiles.
A source profile is the chemical make-up (not the amount, but the fraction of the total) of the
emissions coming from a source; the activity is how those vary over time.
One possible way to clarify this recommendation would be to change the wording as follows,
"Increase efforts to develop methods to collect and measure source profiles at emissions sources,
and the respective tracers in ambient air that uniquely identify those sources." Such a
See
Appendix I
Note 9
See
Appendix I
Note 10
52

-------
recommendation should focus on two areas: 1) organic speciation; and 2) methods with lower
limits of detection for important trace elements.
b)	Identifying and minimizing the uncertainties associated with measuring the organic
fraction of PM1S.
It is important to note that EPA has several significant ongoing efforts that address this topic and
cover both improvements in the methods (sampling and analysis) and development of calibration
and reference standards. EPA's efforts are noted in the OIG's draft report on pages 26 and 27,
where discussions of the Supersites program and ORD's research efforts to improve Speciation
are included. EPA is also developing methods to characterize PM2 5 mass associated with the
organic carbon as measured in the speciation program.
c)	Re-evaluating the methods used in the measurement of ambient ammonia by developing
the proper filter needed to measure PM1S constituents that increase in mass from
absorbing moisture, or, in other instances, the constituents [that] decrease in mass as a
result of volatilization.
We ask that the OIG clarify the statement: "by developing the proper filter needed to
measure PM2 5 constituents that increase in mass from absorbing moisture or, in other
instances, the constituents [that] decrease in mass as a result of volatilization."
-	In the body of the draft report, there is reference to water absorption by
ammonium sulfate. However, if the concern is ammonium sulfate, then a filter will not
make a difference because we use a Teflon filter for mass and it does not absorb water.
-	If the draft report is referring to water associated with ammonium sulfate that possibly
affects the measurement of sulfate, the filter is not an issue as we measure sulfate or
sulfur mass directly and water does not impact the method. However, the water
associated with hygroscopic ammonium sulfate is part of the measured PM2 5 mass as
collected by the Federal Reference Method (FRM) sampler on Teflon filter media. EPA
recognizes that this must be considered when developing control strategies, as it did for
the proposed CAIR.
-	If the report is referring to the measurement of ammonium nitrate, nitrate and
ammonium are measured directly, although there is evidence that ammonium is lost from
nylon filters (4-City Study report). It is also unclear if this is a question about ammonia
or ammonium since the two have been confused in the document. Measurements of
ammonia and nitric acid have not been included in the speciation network. The current
STN collects ions (including ammonium) on a nylon filter and includes a denuder to
remove acid gases (including nitric acid) from the sample stream. Ammonia is not
currently collected using the particle filter, but can be measured using other proven
methods. These gas-phase measurements require different sample collection and analysis
methods. Ammonia and nitric acid gas-phase measurements are being recommended as
part of the EPA National Air Monitoring Strategy NCore level 2 network. EPA
recognizes that ammonium nitrate is semi-volatile, and the amount of particle nitrate that
is part of PM2 5 mass as measured by the FRM is different than the nitrate measured by
See
Appendix I
Note 11
53

-------
the speciation samplers. Methods are available to adjust for this difference for PM
implementation and control strategy development.
-	If the report is referring to the measurement of the precursor ammonia, these
measurements are not done in either the STN or IMPROVE. Methods for ammonia are
well documented in the literature and have been used in monitoring networks for 20+
years. So the question is: does the report refer to the need to measure ammonia properly
in the networks?
-	Reference is also made to the loss of volatile species and a decrease in mass. This does
affect the mass of the "ambient" PM as measured on the Teflon filter, but this is noted in
the FRM Regulations and is accounted for in the PM2 5 standards as the health effects
were measured against mass produced by similar fine particle samplers also using Teflon
filters.
d) Developing and deploying continuous speciation monitors that help provide the real-time
data needed to more accurately depict what is occurring in the atmosphere on a real-time
basis and better pinpoint the sources of PM2,5.
EPA is taking action to address this concern. OAR has deployed a small network of
continuous speciation study sites to aid in the development and implementation of
continuous monitors at routine monitoring sites. This 5-site network has served the needs
well in evaluating the operation and feasibility of the currently available continuous
sulfate, nitrate and carbon monitors in a routine monitoring setting. The State participants
in the study, along with EPA and the vendors, have used this study to help identify issues with
the new monitoring technologies and improve them. OAR plans to expand this study to about 12
sites over the next 2 years, and include newly available continuous speciation monitors. As the
new technologies are demonstrated for use in a routine setting, these sites will serve as the
platform for the long-term continuous monitoring network.
3-4 Establish a stakeholders workgroup to address the challenges described in
Recommendations 3-1, 3-2, and 3-3, comprised of officials from OA OPS, ORD, and selected
EPA Regions; State, local, and tribal agencies; State and Territorial Air Pollution Program
Administrators/Association of Local Air Pollution Control Officials; RPOs; affected
industries; academia; and monitor manufacturers.
In light of the many coordination and advisory processes already in place, we do not
support the recommendation for a new workgroup. We acknowledge and value
participation, feedback and input from our stakeholders, scientific experts, and air
monitoring experts. Our current and upcoming mechanisms for soliciting input provides
for better decision making and program improvement and development. OAR has access
to the newly formed CASAC Ambient Air Monitoring and Methods Subcommittee. This
subcommittee has representatives from State and local government agencies and academia. OAR
is also in the process of forming an ambient air monitoring steering committee composed of
EPA's ORD and OAR, EPA Regional offices, State, local and tribal agencies, and other Federal
agencies. The CASAC subcommittee has recently reviewed the National Air Monitoring
Strategy. The CASAC meetings are open to the public and have involved industry and the
See
Appendix I
Note 12
See
Appendix I
Note 13
54

-------
manufacturers. The combination of these two groups can be used to effectively vet ambient air
monitoring issues and get sufficient and informed feedback on our plans to address challenges.
3-5 Through the workgroup discussed in Recommendation 3-4, increase partnering efforts
with monitor manufacturers to maximize the availability and use of current continuous
speciation monitors and expedite the development of the next generation of speciation
monitors to address the challenges described above. Given the health and economic
consequences if controls are not implemented expeditiously and at the right sources, EPA
should consider a joint EPA-private sector pre-competitive technological research program
similar to the groundbreaking Partnership for a New Generation of Vehicles (PNGV)
program that helped to develop a new generation of low emitting vehicles.
EPA agrees with the intent of this recommendation. Improvements in communication with the
vendor community add value to the development and implementation of current and future
generations of continuous monitors. OAR has continually communicated with the vendors about
monitoring needs and future directions. For example, we have a continuous monitoring study
that requires us to keep in communication with the vendors to present issues, work with them on
resolutions, and implement the latest version of their monitoring technologies. We have been
open about the number of monitoring sites we anticipate and clear that we cannot recommend a
specific vendor type. In contrast to the PNGV, the market for monitoring equipment is quite
limited, so we respectfully disagree that PNGV is a suitable conceptual model for OAR's efforts
on monitoring technology.
It is very important to recognize that EPA must be careful in establishing partnering
relationships with monitoring vendors. Generally, the vendors are looking for some type
of commitment from EPA, either to provide resources or to deploy methods in national
monitoring networks. EPA must be extremely cautious about making such commitments,
and in some cases, will not be able to do so, particularly with respect to recommending a
specific vendor's instrument. ORD's Small Business Innovative Research (SBIR) program is
another program that could be potentially be utilized and, in fact, has been utilized to address
continuous PM mass technologies.
See
Appendix I
Note 14
Suggested Changes to the Text of the Report
In the section, At a Glance, under What We Found: Please revise the 2nd through 4th
sentences related to insufficiency of the speciation data to effectively develop control
strategies. As written, they are incorrect. Suggest revising the text as follows:
See
Appendix I
Note 15
"Although the speciation network provides information on understanding the make-up and
origin of PM2 5, the Agency's ambient monitoring network does not by itself provide the data
needed for EPA or States to identify or quantify the chemical make-up of PM2 5 particles, reliably
trace particles back to their source, or account for chemical changes that occur after particles are
released into the atmosphere. The development of control strategies is best approached through
collaborative processes that use emissions inventories, ambient monitoring data, and air quality
modeling. Speciation data is available to begin work on developing control strategies. EPA and
the States are in the process of using the available monitoring data from the Speciation,
Supersites, and other state and private monitoring networks to begin development of control
strategies; however, increased efforts are needed."
55

-------
In the section, At a Glance, under What We Recommend: Please consider revising the 2nd
sentence of the 1st paragraph to read: "This would include promoting greater attention to
providing opportunities for cooperation with the private sector to develop improved
continuous speciation monitors."
Page 1, 2nd paragraph: . .how the particle can be traced to its source of origin, also known as
fingerprinting;..." Suggest change the wording "also known as fingerprinting" to "through the
use of source apportionment modeling"
Page 11, 3rd bullet: Change the reference to "ammonia" instead of "ammonium". The Speciation
program provides a measure of particulate ammonium but not gas-phase ammonia.
Page 16, 2nd paragraph: "The current state of scientific understanding on the formation of
secondary organic aerosols is insufficient, and as a result PM modeling predictions at the present
time have substantial uncertainties. Improved speciation data would help decrease these
limitations." Suggest clarifying the data needs to support PM modeling predictions. If this is
continuous or semi-continuous speciation data, then this should be clarified in the text.
Page 17: "Key Agency officials agreed that continuous speciation monitors would be the most
likely approach to providing the robust data set needed." Insert the words "or semi-continuous"
after continuous. Also include this text, "semi-continuous monitors for speciation are available
for carbon, nitrate and sulfate. These monitors have the ability to provide more time resolved
data."
Page 18, 2nd paragraph, 1st sentence: change "ammonia" to "ammonium"
Page 21, 2nd paragraph: There are issues with this and the next 2 paragraphs regarding the
discussion of ammonia versus ammonium. The statement: ".. .ammonia is more complicated
because the nylon filter does not bond with ammonia..ammonia is the gaseous, not the
particle species. This discussion is confusing and needs clarification regarding the
appropriateness of particle ammonium measurements in the speciation network and the need for
supplemental gas-phase measurements of ammonia. Please contact Joann Rice in OAQPS, at
919-541-3372 for assistance in clarification.
Page 21, last paragraph, last sentence: "According to an ORD official, measurements of
ammonia and nitric acid, while desired, have not been included in the network due to operational
resources and cost." Please either delete the sentence or include the following statements for
clarification: "These are gaseous, not particle species, and therefore cannot be obtained from
particle filter measurements made by the Speciation network and require different sample
collection and analysis methods. However, the NAAMS NCore Level 2 sites include plans to
include ammonia and nitric acid measurements as part of the multi-pollutant strategy."
Page 22, 2nd paragraph: ".. .there are concerns that without improved speciation monitoring data
on carbon, ammonia, ..." change "ammonia" to ammonium. For clarification, a sentence could
be added that expresses the need for gas-phase measurements of ammonia. Similar issues exist
with the use of the word "ammonia" on page 24, 1st paragraph and in recommendations 3-lb) and
3-3c) starting on page 30. Please change these to "ammonium".
See
Appendix I
Note 16
56

-------
Appendix I
OIG Evaluation of Consolidated EPA Response
to Draft Report
Note 1 - We agree that the Agency has enough information to begin the development of
control strategies, as evidenced in the NARSTO report that states that research
initiated 5 to 10 years ago provides a basic understanding of PM formation,
transport, and its major contributing sources. In addition, we modified the report
to further emphasize that monitoring data assists in the development of an
effective control strategy. However, we continue to believe that increased efforts
are needed to ensure that the States have the data needed to reduce harmful levels
of PM2 5 expeditiously and at the least cost to industry. EPA's response (page 46)
similarly cites the need for improved speciation data, stating that speciated data is
valuable in crafting control strategies to address the principal sources of PM
problems, as well as to assist in better understanding the components of PM that
are of greatest significance to human health effects. Further, key Agency officials
agreed that EPA needed to increase funding from 5 to 10 percent for performing
analytical assessments, adopting new methods, and conducting research on
technologies that can more fully assist in identifying the chemical make-up of
PM2 5, account for the atmospheric impacts on PM2 5, and assay the resultant
changes that occur to the composition of the particle.
Note 2 - As shown in section entitled Speciation Data Used to Groundtruth Emissions
Inventory Estimates and Modeling Assumptions (page 13), our draft report already
noted that monitoring data is one of the three interdependent tools for managing
PM2 5 programs. We agree that all three tools are important to developing an
effective control strategy. We do believe, however, that monitoring data, although
perhaps no more important than emissions estimates and modeled assumptions,
does provide more reliable, and thus more useful, data and, in this context, is used
to groundtruth the other estimates and assumptions.
Note 3 - In Appendix F, we further defined the Supersites Program. However, we believe
the report sufficiently characterizes the role of the Supersites Program, as well as
other EPA activities that support EPA's PM program. For example, page 20 of
the report mentions several EPA efforts underway, such as the STAR program,
ORD's extramural research grants program, and the analytical laboratory
dedicated to measuring organic compounds in PM. Also, because the eight
Supersites are not a nationwide network, we do not agree that the Supersites
Program addresses our recommendations.
Note 4 - We agree with EPA's description of the roles of the Speciation program and that it
will be valuable in developing control strategies and in better understanding
health effects. This is precisely why we believe improvements are needed in the
Speciation program, as defined in the report's recommendations.
57

-------
Note 5 - We agree that EPA has implemented controls and reduced PM levels, as described
in the section entitled Other Programs Impact PM25 Levels on page 10 of the
report, which describes the programs and national control strategies EPA credits
with these reductions. We also agree that, currently, it is difficult for EPA to say
what is and is not an effective control strategy, which suggests that more
information such as speciation data is needed. Finally, we also agree that more
speciation data is needed to identify and control the PM sources posing the
greatest health risk.
Note 6- We agree that the Speciation Monitoring program is only one of three tools
needed for PM NAAQS implementation, the other two being emissions
characterization and air quality process understanding and modeling. We also
understand that the Agency must balance EPA's needs for research not only
within the air research program, but also across all environmental activities. We
are not recommending a major shift in resources to monitoring at the expense of
emissions and modeling, but we do maintain that some level of increased effort is
needed, and continue to believe that a 5-percent increase is appropriate. (Also see
Note 1.)
Note 7 - We revised recommendation 3-1 (a) to more specifically indicate that EPA should
focus its efforts in the area of continuous speciation monitoring, which is also
stated by EPA in its response to the draft report as being a need. We agree that
efforts should include improved methods that will reduce the uncertainties in the
measurement methods. Also, we did not recommend that EPA increase speciation
sampling for pollutants in parts of the nation where reliable emissions information
indicates there are few or no sources.
Note 8 - EPA stated that it is capable of providing a broad characterization of sources
contributing to increased PM2 5 levels. However, we believe that, as State and
local agencies begin to require specific industrial sources to install expensive
controls, a more narrow characterization will be needed. EPA expressed this
viewpoint earlier in its response when it stated that speciated data will assist in
better understanding the components of PM that are of greatest significance to
human health effects, which would improve the input data needed to narrow
characterizations. As our report notes, EPA still has time to overcome these
challenges, but increased efforts will be needed.
Note 9 - In the section entitled Supplementing STN With IMPROVE Data Provides Some
Useful Information But Compatibility Is Limited, we recognize EPA's efforts in
assessing the comparability of STN and IMPROVE.
Note 10 - The recommendation mentions both source profiling and tracer species because
both can impact source apportionment. We agree with EPA's comment and have
modified the recommendation to be more specific as suggested in EPA's
response.
Note 11 - We agree, and have revised the report and recommendation to reflect the
Agency's comments by removing reference to developing the proper filter.
58

-------
Note 12 - We agree that the Agency is taking actions to address this recommendation;
however, we believe increased efforts are needed. In its response, EPA supports
increased effort by stating that improvements in real-time measurements through
increased use of continuous monitors will improve the Agency's ability to more
specifically identify sources of PM.
Note 13 - We agree that the two existing workgroups can effectively address
Recommendations 3-1, 3-2, and 3-3, provided the Agency implements these
recommendations by addressing the issues cited in this report as part of the work
carried out by either the steering committee or the Clean Air Scientific Advisory
Committee's Ambient Air Monitoring and Methods Subcommittee.
Note 14 - We agree that the Agency must be careful in establishing partnering relationships
with monitoring vendors, and be certain to maintain its independence. However,
the monitor manufacturers' willingness to invest their own resources in research
and development of new and improved monitoring equipment is an important
resource for the development and improvement of the next generation of
speciation monitors. Further, we do not see the dissimilarities with EPA's efforts
to partner with the major auto manufacturers under PNGV, and our
recommendation that the Agency partner with monitoring manufacturers under a
similar approach.
Note 15 - We generally agree with the Agency's suggested revisions and have modified the
report where appropriate. However, the speciation monitoring network's data
currently are not sufficient to assist EPA and the States in fully tracing particles
back to their source, or accounting for chemical changes that occur after particles
are released into the atmosphere.
Note 16 - For the remainder of EPA's response under the section entitled Suggested
Changes to the Text of the Report, we agree with the Agency's comments and
have revised the report as appropriate.
59

-------
Appendix J
Distribution
EPA Headquarters
Assistant Administrator for Air and Radiation (6101 A)
Deputy Assistant Administrator for Science, Office of Research and Development (8105R)
Deputy Assistant Administrator for Air and Radiation (6101 A)
Agency Followup Official (the CFO) (271 OA)
Agency Followup Coordinator (2724A)
Audit Followup Coordinator, Office of Air and Radiation (6102 A)
Associate Administrator for Congressional and Intergovernmental Relations (1301 A)
Associate Administrator for Public Affairs (1101 A)
Director, Office of Air Quality Planning and Standards (C404-04)
Deputy Director, Office of Air Quality Planning and Standards (C404-04)
Director, Emissions Standards Division (C504-03)
Acting Director, Emissions, Monitoring and Analysis Division (C304-02)
Director, National Exposure Research Laboratory (MD-75)
Audit Liaison, Office of Air Quality Planning and Standards (C404-2)
Audit Liaison, Office of Research and Development (8102R)
EPA Regions
Regional Air Program Directors
EPA Office of Inspector General
Inspector General (2410)
60

-------