Policy Assessment
for the Review of the Participate Matter
National Ambient Air Quality Standards
First External Review Draft

March 2010

-------
                                    DISCLAIMER
       This draft document has been prepared by staff from the Ambient Standards Group,
Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency. Any
opinions, findings, conclusions, or recommendations are those of the authors and do not
necessarily reflect the views of the EPA. This document is being circulated to obtain review and
comment from the Clean Air Scientific Advisory Committee (CASAC) and the general public.
Comments on this draft document should be addressed to Beth Hassett-Sipple, U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, C504-06,
Research Triangle Park, North Carolina 27711 (email: hassett-sipple.beth@epa.gov).

-------
                                          EPA 452/P-10-003
                                               March 2010
            Policy Assessment
for the Review of the Paniculate Matter
National Ambient Air Quality Standards
          First External Review Draft
          U.S. Environmental Protection Agency
              Office of Air and Radiation
        Office of Air Quality Planning and Standards
        Research Triangle Park, North Carolina 27711

-------
This page left intentionally blank.

-------
                          TABLE OF CONTENTS
1  INTRODUCTION	1-1
   1.1    PURPOSE	1-1
   1.2    BACKGROUND	1-2
      1.2.1    Legislative Requirements	1-2
      1.2.2    Overview of theNAAQS Review Process	1-3
      1.2.3    History of PMNAAQS Reviews	1-5
      1.2.4    Litigation Related to the 2006 PM Standards	1-10
      1.2.5    Current PM NAAQS Review	1-11
   1.3    GENERAL APPROACH AND ORGANIZATION OF THIS DOCUMENT	1-12
   1.4    REFERENCES	1-14
2  REVIEW OF THE PRIMARY STANDARDS FOR FINE PARTICLES	2-1
   2.1    APPROACH	2-2
      2.1.1    Approaches Used in Previous Reviews	2-3
            2.1.1.1     Review Completed in 1997	2-3
            2.1.1.2     Review Completed in 2006	2-5
      2.1.2    Remand of Primary Annual PM2.5 Standard	2-8
      2.1.3    Current Approach	2-10
   2.2    ADEQUACY OF CURRENT STANDARDS	2-14
      2.2.1    Evidence-based Considerations	2-15
      2.2.2    Risk-based Considerations	2-43
      2.2.3    Preliminary Staff Conclusions on Adequacy of Current Standards	2-55
   2.3    CONSIDERATION OF ALTERNATIVE STANDARDS	2-56
      2.3.1    Indicator	2-56
      2.3.2    Averaging Times	2-66
      2.3.3    Forms	2-71
            2.3.3.1     Form of the Annual Standard	2-71
            2.3.3.2     Form of the 24-Hour Standard	2-73
      2.3.4    Alternative Levels to Address Health Effects Related to Long-term PM2.5
              Exposures	2-74
            23.4.1     Evidence-based Considerations	2-74
            2.3.4.2     Risk-based Considerations	2-81

      DRAFT DO NOT QUOTE OR CITE      i                            MARCH 2010

-------
      2.3.5   Alternative Levels to Address Health Effects Related to Short-term PM2.5
             Exposures	2-88
            2.3.5.1     Evidence-based Considerations	2-89
            2.3.5.2     Risk-based Considerations to Inform Alternative Levels	2-99
      2.3.6   Preliminary Staff Conclusions on Alternative Levels to Address Health Effects
             Related to Long- and Short-term PM2.5 Exposures	2-104
   2.4    PRELIMINARY STAFF CONCLUSIONS ON ALTERNATIVE FINE PARTICLE
         STANDARDS	2-108
   2.5    KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH AND DATA
         COLLECTION	2-110
   2.6    REFERENCES	2-111
3  REVIEW OF THE PRIMARY STANDARD FOR THORACIC COARSE
   PARTICLES	3-1
   3.1    APPROACH	3-1
      3.1.1    Approaches Used in Previous Reviews	3-2
      3.1.1.1         Reviews Completed in 1987 and 1997	3-2
      3.1.1.2         Review Completed in 2006	3-3
      3.1.2    Litigation of 2006 Final Rule for thoracic coarse particles	3-5
      3.1.3    Approach in the Current Review	3-7
   3.2    ADEQUACY OF THE EXISTING PMio STANDARD	3-8
      3.2.1    Evidence of Effects Related to Ambient Thoracic Coarse Particles	3-9
      3.2.2    Preliminary Staff Conclusions on Adequacy of Current PMio Standard	3-22
   3.3    CONSIDERATION OF POTENTIAL ALTERNATIVE STANDARDS	3-25
      3.3.1    Indicator	3-26
      3.3.2    Consideration of Indicator in the Review Completed in 2006	3-26
      3.3.3    Evidence-Based Consideration of Indicator in the Current Review	3-31
      3.3.4    Averaging Time	3-34
      3.3.5    Level and Form	3-35
      3.3.6    Form	3-37
      3.3.7    Level	3-37
   3.4    KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH AND DATA
   COLLECTION	3-39
   3.5    REFERENCES	3-39
4  REVIEW OF THE SECONDARY STANDARDS FOR VISIBILITY-RELATED
   EFFECTS	4-1

      DRAFT DO NOT OUOTE OR CITE     ii                           MARCH 2010

-------
   4.1    APPROACH	4-2
      4.1.1    Approaches Used in Previous Reviews	4-2
      4.1.2    Remand of Secondary PM2.5 Standards	4-8
      4.1.3    Current Approach	4-9
   4.2    ADEQUACY OF CURRENT STANDARDS	4-10
      4.2.1    Evidence-based Considerations	4-10
      4.2.2    Summary	4-28
   4.3    CONSIDERATION OF ALTERNATIVE STANDARDS	4-29
      4.3.1    Nature of the Indicator	4-30
      4.3.2    Averaging and Applicable Times	4-32
      4.3.3    Alternative Levels/Forms	4-33
      4.3.4    Performance of Alternative Standards	4-35
   4.4    PRELIMINARY STAFF CONCLUSIONS ON ALTERNATIVE
         SECONDARY PM STANDARDS FOR VISIBILITY-RELATED
         EFFECTS	4-42
   4.5    REFERENCES	4-45
   APPENDIX 4A. INFORMATION REGARDING THE 1-HOUR PM2.5 MASS
   INDICATOR	A-l
5  REVIEW OF THE SECONDARY STANDARDS FOR OTHER WELFARE
   EFFECTS	5-1
   5.1    APPROACH	5-2
      5.1.1    Approaches Used in Previous Reviews	5-2
      5.1.2    Scope of Current NAAQS Reviews	5-3
      5.1.3    Current Approach	5-6
   5.2    CLIMATE	5-6
      5.2.1    Scope	5-7
      5.2.2    Adequacy of the Current Standard	5-8
      5.2.3    Preliminary Staff Conclusions	5-11
      5.2.4    Key Uncertainties and Areas for Future Research and Data Collection	5-12
   5.3    ECOLOGICAL EFFECTS	5-13
      5.3.1    Scope	5-13
      5.3.2    Adequacy of the Current Standard	5-16
      5.3.3    Preliminary Staff Conclusions	5-23


      DRAFT DO NOT QUOTE OR CITE     iii                          MARCH 2010

-------
   5.3.4     Key Uncertainties and Areas for Future Research and Data Collection	5-24
5.4   MATERIALS	5-25
   5.4.1     Scope	5-25
   5.4.2     Adequacy of the Current Standard	5-25
   5.4.3     Preliminary Staff Conclusions	5-27
   5.4.4     Key Uncertainties and Areas for Future Research and Data Collection	5-28
5.5   REFERENCES	5-28
   DRAFT DO NOT QUOTE OR CITE     iv                            MARCH 2010

-------
 1                                    LIST OF TABLES
 2
 3    Table 1-1.  Summary of National Ambient Air Quality Standards Promulgated for Particulate
 4              Matter 1971-2006	1-6
 5    Table 2-1.  Summary of Causality Determinations for PM2.5	2-16
 6    Table 2-2.  Estimated Incidence and Percent of Total Annual Incidence Associated with
 7              Long-term PM2.s Exposure Based on Simulation of the Current Suite of Standards
 8              (for fflD mortality based on 2007 PM2.5 Concentrations)1'2	2-46
 9    Table 2-3. Estimated Incidence and Percent of Total Annual Incidence Associated with Short-
10              Term PM2.s Exposure Based on Simulation of the Current Suite of Standards (CV
11              mortality and hospital admissions based on 2007 PM2.s concentrations)1'2	2-50
12    Table 2-4.  Comparison of Air Quality Data for Selected Epidemiological Studies -
13              Long-term Mean Concentrations	2-79
14    Table 2-5.  Comparison of Air Quality Data for Selected Epidemiological Studies- 98 Percentile
15              Values	2-94
16    Table 3-1.  Summary of Causality Determinations for PMio-i.s	3-11
17    Table 3-2.  PMi0 Concentrations in Cities with Statistically Significant PMi0-2.5 Effect
18              Estimates	3-18
19    Table 4-1 Logit model estimated VAQ values corresponding to various percent acceptability
20              values for the four cities	4-20
21    Table 4-2 Percentage of daily maximum hourly values of daylight PM light extinction
22              exceeding  CPLs (excluding hours with relative humidity greater than 90 percent).
23              (adapted from Table 3-7 in UFVA)	4-25
24    Table 4-3 Percentage of daily maximum hourly values of daylight PM light extinction
25              exceeding  CPLs when "just meeting" the current PM2.5 NAAQS (15/35 ug/m3)
26              (excluding hours with relative humidity greater than 90 percent).
27              (Adapted from Table 4-7 in UFVA)	4-28
28    Table 4.3 1. PM light extinction design values for "just meeting" secondary NAAQS scenarios
29              based on measured PM light extinction (excluding hours with relative humidity
30              greater than 90 percent)	4-38
31    Table 4.3-2 90th percentile maximum daily  1-hour PM light extinction design values (Mm-1)
32              after rollback to meet alternative standard of 60 ug/m3, 40 ug/m3, 30 ug/m3,20
33              ug/m3 and 10 ug/m3 maximum daylight 1-hour PM mass concentration
34              for the  90th percentile.*	4-41
35    Table A-l. Percentage of days with daily maximum daylight 1-hour PM2.5 mass concentration
36              exceeding  reference levels in 2005-2007 (excluding hour with relative humidity
37              greater than 90 percent)	A-3
38    Table A-2. Alternative NAAQS scenarios based on daily maximum daylight 1-hour PM2.5
39              mass, averaged over three years (excluding hours with relative  humidity greater than
40              90 percent)	A-4
41    Table A-3. 2005-2007 design values for 1-hour PM2.5 mass (|ig/m3 )	A-7
42    Table A-4. Percentage reductions in non-PRB PM2.5 components required to meet NAAQS
43              scenarios based on 1-hour PM2.5 mass	A-7

-------
 1    Table A-5. Post-rollback design values for daily maximum 1-hour PM2.5 mass. Design
 2              values are shown only for combinations of study area and scenario for which the
 3              study area does not meet the scenario under current conditions, such that reductions
 4              were made during the rollback modeling	A-9
 5    Table A-6. Percentage of days across three years (two in the case of Phoenix and Houston)
 6              with maximum 1-hour daylight PM light extinction above CPLs when "just meeting"
 7              NAAQS scenarios based on 1-hour PM2.5 mass. Blue shading indicates no
 8              reduction required from current conditions	A-16
 9    Table 5-1.  Scope of the current secondary PM NAAQS review and current NOx/SOx secondary
10              review	5-4
11
12                                    LIST OF FIGURES
13
14    Figure 2-1. Summary of Effect Estimates (per 10 |ig/m3) and Air Quality Distributions for
15              Long-term Exposure Studies	2-23
16    Figure 2-2. Distribution of study-specific area mean PM2.5 concentrations	2-75
17    Figure 2-3. Percent reduction in long-term exposure-related mortality risk	2-82
18    Figure 2-4  Summary of Effect Estimates (per 10 |ig/m3) - Short-term Exposure Studies	2-91
19    Figure 2-5. Distribution of Study-specific 98th Percentile Concentrations	2-95
20    Figure 2-6. Percent Reduction in Short-term Exposure-related Mortality and
21              Morbidity Risk	2-100
22    Figure 4-1  Summary of results of urban visibility studies in four cities, showing the identified
23              range of the 50% acceptance criteria  	4-19
24    Figure 4 2. Distribution of estimated maximum daily daylight 1-hour PM light extinction across
25              the 2005-2007 period, by study area (excluding hours with relative humidity greater
26              than 90 percent). (Adapted from Figure 3-8 in UFVA)*	4-24
27    Figure 4 3. Distribution of daylight 1 -hour PM light extinction when rolled back to just meet
28              current PM fine NAAQS across the 2005-2007 period, by  study area (excluding
29              hours with relative humidity greater than 90 percent)	4-27
30    Figure 4.3-1 Scatter plots of PM light extinction versus PM2.5 mass concentration of two cities
31              (from UFVA Appendix D, Figure D2)	4-32
32    Figure 4.3-2 Daily maximum daylight 1-hour light extinction after rollback to just meet a
33              scenario with daily maximum of 112 Mm"1 for the 90th percentile excluding hours
34              with relative humidity greater than 90 percent	4-37
35    Figure 4.3-3  Maximum daily daylight 1-hour PM light extinction under "just meet"  conditions
36              for a NAAQS scenario based on  1-hour PMi.s mass of 30 ug/m3, 90th percentile,
37              excluding relative humidity >90%	4-39
38    Figure A-l. 2005-2007 daily maximum daylight 1-hour PM2.5 mass concentrations (|ig/m3)
39              for the 15 study areas (excluding hours with relative humidity greater than 90
40              percent)	A-2
41    Figure A-2. Distributions of maximum daily daylight 1-hour PM light extinction under "just
42              meet" conditions for NAAQS scenarios based on 1-hour PM2.5 mass (excluding
43              hours >90% RH)	A-ll

-------
1 L
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
1ST OF ACR

ACS
AHR
AHSMOG
ANS
AQCD
AQS
AR4

BC
BenMAP
BP
C
Ca
CAA
CAPs
CARS
CASAC
CBS A
CBVD
CCN
CCSP
Cd
CHD
CHF
CHS
CO
COPD
C-R
CRP
CSN
CTM
Cu
American Cancer Society
Airway Hyperesponsiveness
California Seventh Day Adventist Study
Autonomic nervous system
Air Quality Criteria Document
EPA's Air Quality System
Fourth Assessment Report of the Intergovernmental Panel on Climate
Change
Black carbon
Benefits Mapping Analysis Program
Blood pressure
Carbon
Calcium
Clean Air Act
Concentrated Ambient Particles
California Air Resources Board
Clean Air Scientific Advisory Committee
Consolidated Business Statistical Area
Cerebrovascular disease
Cloud Condensation Nuclei
US Climate Change Science Program
Cadmium
Coronary Heart disease
Congestive heart failure
Childrens Health Study
Carbon Monoxide
Chronic obstructive pulmonary disease
Concentration-response relationship
C-reactive protein
Chemical Speciation Network
Chemical transport models
Copper

-------
1
2
O
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
CV
CVD
DE
DEP
dv
DVT
EC
ECG
ED
EPA
FEM
FEVi
FRM
GAMs
GEOS
GHG
GI
GLMs
GSH
GST
HA
HEI
Hg
HR
HRV
ICD
ICR
IFG
IHD
IMPROVE
IPCC
IRP
ISA
IT
Cardiovascular
Cardiovascular disease
Diesel Exhaust
Diesel Exhaust Particles
deciview
Deep Vein Thrombosis
Elemental Carbon
Electrocardiogram
Emergency department
Environmental Protection Agency
Federal Equivalent Method
Change in forced expiratory volume in one second
Federal Reference Method
Generalized additive models
Global Scale Air Circulation Model
Greenhouse Gas
Group Interviews
Generalized linear models
Glutathione
Glutathione-S-transferase
Hospital admissions
Health Effects Institute
Mercury
Heart rate
Heart rate variability
International Classification of Disease
Information Collection Request
Investigative Focus Groups
Ischemic heart disease
Interagency Monitoring of Protected Visual Environment
Intergovernmental Panel on Climate Change
Integrated Review Plan
Integrated Science Assessment
Intratracheal

-------
1
2
O
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Km
Lag
LagO

Lag 0-x


MCAPS
MEA
MENTOR
MI
Mm
MOA
MSA
N
NAAQS
NCEA
NCore
Ni
NMMAPS
NO
NO2
MV
NOx
NFS
NRC
NWS
03
OAQPS
OAR
OC
OR
ORD
OS
PA
Kilometer
Time between one event and another
Same day as the death, test, hospital, ED, clinic, physician visit;
that occurs on the same day as the exposure to the pollutant(s)
All the deaths test, hospital, ED, clinic, physician visit; that occurs
on the same day as the exposure to the pollutant(s) and the x days
following the day of exposure
Medicare Air Pollution Study
Millennium Ecosystem Assessment
Modeling ENvironment for TOtal Risk
Myocardial infarction
Megameter
Mode(s) or mechanism(s) of action
Metropolitan Statistical Area
Nitrogen
National Ambient Air Quality Standards
National Center for Environmental Assessment
National Core Monitoring Network
Nickel
National Morbidity, Mortality, and Air Pollution Study
Nitric Oxide
Nitrogen dioxide
Nitrate
Nitrogen oxides
National Park Service
National Research Council
National Weather Service
Ozone
Office of Air Quality Planning and Standards
Office of Air and Radiation
Organic Carbon
Odds Ratio
Office of Research and Development
Observational Study
Policy Assessment

-------
 1          PA                 Pulmonary arterial
 2          PAH               Polyaromatic Hydrocarbon
 3          Pb                  Lead
 4          PEF                Peak Expiratory Flow L/min
 5          PM                 Particulate matter
 6          PMio               Particles with an upper 50% cut-point of 10± 0.5 um aerodynamic
 7                              diameter and a penetration curve as specified in the Code of
 8                              Federal Regulations.
 9          PMio-2.5             Particles with an upper 50% cut-point of 10 um aerodynamic
10                              diameter and a lower 50% cut-point of 2.5 um aerodynamic
11                              diameter.
12
13          PM2.5               Particles with an upper 50% cut-point of 2.5 um aerodynamic
14                              diameter and a penetration curve as specified in the Code of
15                              Federal Regulations.
16
17          PMx               The legal definition for PMx,  as defined in the Code of Federal
18                              Regulations, includes both a 50% cut-point and a penetration
19                              curve.  A 50% cut-point of X  um diameter means that 50% of
20                              particles with aerodynamic diameter of X are removed by the inlet
21                              and 50% pass through the inlet and are collected on the filter.
22                              Depending on the specific penetration curve specified, particles
23                              larger than X um aerodynamic diameter are collected with an
24                              efficiently than decreases rapidly for particles larger than X while
25                              the collection efficiency for particles smaller than X increases
26                              rapidly with decreasing size until 100 % efficiency is reached.
27          PRB                Policy-Relevant Background
28          PSAS               The French National Program on Air Pollution Health Effects
29          PT                  Prothrombin Time
30          PTT                Partial Thomboplastin Time
31          QA                 Quality assurance
32          QT                 Time for depolarization and repolarization of the ventricles
33          RA                 Risk Assessment
34          RE A               Risk and Exposure Assessment
35          RF                  Radiative forcing
36          RH                 Relative humidity
37          RNS               Reactive Nitrogen Species
38          ROS               Reactive Oxygen Species
39          RR                 Relative risk

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
SAB
SANDWICH

SAP
SBP
SD
SEARCH
SEDD
SES
Si
SID
SMOKE
S
S02
SO42"
sox
SOPHIA
STP
TB
TSP
UFPs
UFVA
V
VAQ
VOC
WACAP
WHI
Zn
Science Advisory Board
Sulfate, Adjusted Nitrate, Derived Water, Inferred Carbonaceous
mass approach
Synthesis and Assessment Product
Systolic Blood Pressure
Standard deviation
Southeastern Aerosol Research and Characterization Study
State Emergency Department Databases
Socioeconomic Status
Silicon
State Inpatient Database
Sparse Matrix Operator Kernal Emissions
Sulfur
Sulfur Dioxide
Sulfate
Sulfur Oxides
Study of Particulates and Health in Atlanta
Standard Temperature and Pressure
Tracheobronchial
Total  suspended particulate
Ultrafine particles
Urban-Focused Visibility Assessment
Vanadium
Visual Air Quality
Volatile organic compounds
Western Airborne Contaminants Assessment Project
Women's Health Initiative
Zinc

-------
1
2                                   This page intentionally left blank.

-------
 1                                    1. INTRODUCTION

 2    1.1     PURPOSE
 3           The U.S. Environmental Protection Agency (EPA) is presently conducting a review of
 4    the national ambient air quality standards (NAAQS) for particulate matter (PM).  The overall
 5    plan and schedule for this review were presented in the Integrated Review Plan for the National
 6    Ambient Air Quality Standards for Particulate Matter (IRP; US EPA, 2008a). The IRP
 7    identified key policy-relevant issues to be addressed in this review as a series of questions that
 8    frame our consideration of whether the current NAAQS for PM should be retained or revised.
 9           This Policy Assessment (PA), prepared by staff in the EPA's Office of Air Quality
10    Planning and Standards (OAQPS), is intended to help "bridge the gap" between the relevant
11    scientific information and assessments and the judgments required of the EPA Administrator in
12    determining whether, and if so, how, it is appropriate to revise the NAAQS for PM.1 This first
13    draft PA presents factors relevant to EPA's review of the primary (health-based) and secondary
14    (welfare-based) PM NAAQS. It focuses on both evidence- and risk-based information in
15    evaluating the adequacy of the current PM NAAQS and in identifying potential alternative
16    standards for consideration. In this first draft PA, we consider the scientific and technical
17    information available in this review as assessed in the Integrated Science Assessment for
18    Particulate Matter (Final Report) (ISA, US EPA, 2009a), the Quantitative Health Risk
19    Assessment for Particulate Matter - Second External Review Draft (US EPA, 201 Oa) and the
20    Particulate Matter Urban-Focused Visibility Assessment - Second External Review Draft (US
21    EPA, 201 Ob). In so doing, we focus on information that is most pertinent to evaluating the basic
22    elements of NAAQS:  indicator2, averaging time, form,3 and level. These elements, which
23    together serve to define each standard, must be considered collectively in evaluating the health
24    and welfare protection afforded by the PM standards.
25           We also recognize that part of the definition of the NAAQS includes specifying
26    allowable monitoring methods by which the indicator is to be measured as well as minimum
27    requirements for monitoring, such as monitor siting criteria.  Such monitoring issues were
28    identified in the IRP (US EPA, 2008a, chapter 7) and we plan to include a discussion of them in
29    the second draft PA.
      1 Preparation of a PA by OAQPS staff reflects Administrator Jackson's decision to modify the NAAQS review
      process that was presented in the IRP.  See http://www.epa.gov/ttn/naaqs/review.html for more information on the
      current NAAQS review process.
      2 The "indicator" of a standard defines the chemical species or mixture that is to be measured in determining
      whether an area attains the standard.
      3 The "form" of a standard defines the air quality statistic that is to be compared to the level of the standard in
      determining whether an area attains the standard.
      Draft Do Not Quote or Cite               1-1                             March 2010

-------
 1           Although this first draft PA should be of use to all parties interested in this PM NAAQS
 2    review, it is written with an expectation that the reader has familiarity with the technical
 3    discussions contained in the ISA (US EPA, 2009a) and in the draft quantitative risk and visibility
 4    assessment documents (US EPA, 2010a,b).

 5    1.2    BACKGROUND

 6    1.2.1   Legislative Requirements
 7           Two sections of the Clean  Air Act (Act) govern the establishment and revision of the
 8    NAAQS.  Section 108 (42  U.S.C.  section 7408) directs the Administrator to identify and list air
 9    pollutants that meet three specified criteria, including air pollutants "emissions of which, in his
10    judgment, cause or contribute to air pollution which may reasonably be anticipated to  endanger
11    public health and welfare"  and whose "presence ... in the ambient air results from numerous or
12    diverse mobile or stationary sources" and to issue air quality criteria for those that are listed.  Air
13    quality criteria are to "accurately reflect the latest scientific knowledge useful in indicating the
14    kind and extent of all identifiable effects on public health or welfare which may be expected
15    from the presence of [a]  pollutant  in the ambient air . . ." 42 U.S.C. § 7408(b).
16           Section 109 (42 U.S.C. section 7409) directs the Administrator to propose and
17    promulgate "primary" and  "secondary" NAAQS for pollutants for which air quality criteria are
18    issued. Section 109(b)(l) defines a primary standard as one "the attainment and maintenance  of
19    which in the judgment of the Administrator, based on [air quality] criteria and allowing an
20    adequate margin of safety,  are requisite to protect the public health."4 A secondary standard, as
21    defined in Section 109(b)(2), must "specify a level of air quality the attainment and maintenance
22    of which, in the judgment of the Administrator, based on such [air quality] criteria, is requisite to
23    protect the public welfare from any known or anticipated adverse effects associated with the
24    presence of [the]  pollutant  in the ambient air."5
25           The requirement that primary standards include an adequate margin of safety was
26    intended to address uncertainties associated with inconclusive scientific and technical
27    information available at the time of standard setting. It was also intended to provide a  reasonable
28    degree of protection against hazards that research  has not yet identified. Lead Industries
      4 The legislative history of section 109 indicates that a primary standard is to be set at "the maximum permissible
      ambient air level.. . which will protect the health of any [sensitive] group of the population," and that for this
      purpose "reference should be made to a representative sample of persons comprising the sensitive group rather than
      to a single person in such a group." S. Rep. No. 91-1196, 91st Cong., 2d Sess. 10 (1970).
      5 Welfare effects as defined in section 302(h) (42 U.S.C. section 7602(h)) include, but are not limited to, "effects on
      soils, water, crops, vegetation, man-made materials, animals, wildlife, weather, visibility and climate, damage to and
      deterioration of property, and hazards to transportation, as well as effects on economic values and on personal
      comfort and well-being."
      Draft Do Not Quote or Cite               1-2                              March 2010

-------
 1   Association v. EPA, 647F.2d 1130, 1154 (D.C. Or 1980), cert, denied, 449 U.S. 1042 (1980);
 2   American Petroleum Institute v. Costle, 665 F.2d 1176, 1186 (D.C. Cir. 1981), cert, denied, 455
 3   U.S. 1034 (1982); American Farm Bureau Federation v. EPA, 559 F. 3d 512, 533 (D.C. Cir.
 4   2009).  Both kinds of uncertainties are components of the risk associated with pollution at levels
 5   below those at which human health effects can be said to occur with reasonable scientific
 6   certainty. Thus, in selecting primary standards that include an adequate margin of safety, the
 7   Administrator is seeking not only to prevent pollution levels that have been demonstrated to be
 8   harmful but also to prevent lower pollutant levels that may pose an unacceptable risk of harm,
 9   even if the risk is not precisely identified as to nature or degree.
10           In selecting a margin of safety, EPA considers such factors as the nature and severity of
11   the health effects involved, the size of the susceptible population(s) at risk, and the kind and
12   degree of the uncertainties that must be addressed. The selection of any particular approach to
13   providing an adequate margin of safety is a policy choice left specifically to the Administrator's
14   judgment. Lead Industries Association v. EPA, supra, 647 F.2d at 1161-62.
15           In setting standards that are "requisite" to protect public health and welfare, as provided
16   in section 109(b), EPA's task is to establish standards that are neither more nor less stringent
17   than necessary for these purposes.  In so doing, EPA may not consider the costs of implementing
18   the standards. See generally Whitman v. American Trucking Associations, 531 U.S. 457, 471,
19   475-76 (2001).
20           Section 109(d) (1) of the Act requires that "not later than December 31, 1980, and at 5-
21   year intervals thereafter, the Administrator shall  complete a thorough review of the criteria
22   published under section 108 and the national ambient air quality standards . . . and shall make
23   such revisions in such criteria and standards and promulgate such new standards as may be
24   appropriate in accordance with section [108]... and subsection (b) . . ." 42 U.S.C. § 7409(d)(l).
25   Section 109(d)(2) requires that an independent scientific review committee "shall complete a
26   review of the criteria . . . and the national primary and secondary ambient air quality standards . .
27   . and shall recommend to the Administrator any new . . . standards and revisions of existing
28   criteria and standards as may be appropriate . . ." 42 U.S.C. § 7409(d)(2).  Since the early
29   1980's, this independent review function has been performed by the Clean Air Scientific
30   Advisory Committee (CASAC).

31   1.2.2   Overview of the NAAQS Review Process
32           Since completion of the last PM NAAQS review, the Agency has made a number of
33   changes to the process for reviewing the NAAQS. In making these changes, the Agency
34   consulted with CASAC and the public.  This revised process contains four major components:
35   planning, science assessment, risk/exposure assessment, and policy assessment/rulemaking.
     Draft Do Not Quote or Cite              1-3                            March 2010

-------
 1          The planning phase begins with a "kick-off workshop to get input from external and
 2    internal experts and the public regarding policy-relevant issues from the last review and others
 3    that have more recently emerged. The workshop discussions help inform the preparation of an
 4    IRP jointly by OAQPS and Office of Research and Development/National Center for
 5    Environmental Assessment (ORD/NCEA) staff.  A draft IRP is presented for consultation with
 6    CAS AC and for public comment. A final IRP reflects consideration of CAS AC and public
 7    comments together with early guidance from Agency management.  The IRP presents the policy -
 8    relevant questions that will frame the review, the review process and schedule, and descriptions
 9    of the purpose, contents, and approach for developing each of the key documents.
10          The science assessment phase involves the preparation of an ISA by NCEA staff.  The
11    ISA provides a concise evaluation and integration of the policy-relevant science, including key
12    science judgments that are important to inform the design and scope of the risk and exposure
13    assessments. The ISA and its supporting annexes provide a comprehensive assessment of the
14    current scientific literature pertaining to known and anticipated effects on public health and
15    welfare associated with the presence of the pollutant in the ambient air, emphasizing information
16    that has become available since the last review.  The process generally includes production of a
17    first and second draft ISA, which undergo CAS AC and public  review prior to completion of the
18    final IS A.
19          In the risk/exposure assessment phase, OAQPS staff draws upon information and
20    conclusions presented in the ISA to develop quantitative estimates of the risks/exposures for
21    health and/or welfare effects associated with current ambient levels of PM, with levels that just
22    meet the current standards, and with levels that just meet potential alternative standards.  The
23    REAs present methods, key results, observations, and related uncertainties. These assessments
24    begin with preparation of a planning document that discusses the scope and methods planned for
25    use in conducting the quantitative assessments.  Scope and Methods Plans are generally prepared
26    in conjunction with the first draft ISA and presented for consultation with CASAC and for public
27    comment.  Comments received on the Scope and Methods Plans are considered by EPA staff in
28    conducting the analyses and preparing draft REAs. The process generally includes production of
29    first and second draft REAs, which undergo CASAC and public review prior to completion of
30    final REAs.
31          The review process ends with a policy assessment/rulemaking phase.  Recent revisions to
32    process have reinstated the preparation of a Policy Assessment (PA). The PA, like the previous
33    Staff Paper, integrates and interprets the information from the ISA and REAs to provide a
34    transparent staff analysis of the scientific basis for alternative policy options for consideration by
35    the Administrator prior to the issuance of proposed and final rules (Jackson, 2009). One or more
36    drafts of the PA undergo CASAC and public review prior to completion of the final PA.  The PA
      Draft Do Not Quote or Cite              1-4                            March 2010

-------
 1    is intended to facilitate CASAC's advice and recommendations to the Administrator on any new
 2    standards or revisions to existing standards as may be appropriate, as provided for in the CAA.
 3    Following issuance of the final PA, the Agency publishes a proposed rule, followed by a public
 4    comment period. Taking into account comments received on the proposed rule, the Agency
 5    issues a final rule to complete the rulemaking process.

 6    1.2.3   History of PM NAAQS Reviews
 7           The NAAQS for PM that have been promulgated to date are summarized in Table 1-1
 8    and briefly discussed below.  Paniculate matter is the generic term for a broad class of
 9    chemically and physically diverse substances that exist as discrete particles (liquid droplets or
10    solids) over a wide range of sizes, such that the indicator for a PM NAAQS has historically been
11    defined in terms of particle size ranges.
12           The EPA first established NAAQS for PM in 1971 (36 FR 8186), based on the original
13    air quality criteria document (DHEW, 1969).  The reference method specified for determining
14    attainment of the original standards was the high-volume sampler, which collects PM up to a
15    nominal size of 25 to 45 micrometers (|im) (referred to as total suspended particles or TSP). The
16    primary standards (measured by the indicator TSP) were 260 |ig/m3, 24-hour average, not to be
17    exceeded more than once per year, and 75  |ig/m3,  annual geometric mean. The secondary
18    standard was 150 |ig/m3, 24-hour average, not to be exceeded more than once per year.
19    In October 1979, EPA announced the first periodic review of the criteria and NAAQS for PM,
20    and significant revisions to the original standards were promulgated in 1987 (52 FR 24634, July
21    1, 1987). In that decision, EPA changed the indicator for PM from TSP to PMio, the latter
22    including particles with a median aerodynamic diameter6 less than or equal to 10 jim, which
23    delineates thoracic particles (i.e., that subset of inhalable particles small enough to penetrate
24    beyond the larynx to the thoracic region of the respiratory tract).  The EPA also revised the
25    primary standards by: (1) replacing the 24-hour TSP standard with a 24-hour PMio standard of
26    150 |ig/m3 with no more than one expected exceedance per year; and (2) replacing the annual
27    TSP standard with a PMio standard of 50 |ig/m3, annual arithmetic mean.  The secondary
28    standard was revised by replacing it with 24-hour and annual standards identical in all respects to
29    the primary standards. The revisions also included a new reference method for the measurement
30    of PMio in the  ambient air and rules for determining attainment of the new standards. On
      6 The more precise term is 50 percent cutpoint or 50 percent diameter (D50). This is the aerodynamic particle
      diameter for which the efficiency of particle collection is 50 percent.  Larger particles are not excluded altogether,
      but are collected with substantially decreasing efficiency and smaller particles are collected with increasing (up to
      100 percent) efficiency.

      Draft Do Not Quote or Cite               1-5                             March 2010

-------
1         Table 1-1. Summary of National Ambient Air Quality Standards Promulgated for
2                                     Particulate Matter 1971-20067
Final Rule
1971
(36 FR 8186
April 30, 1971)
1987
(52 FR 24634;
July 1, 1987)
1997
(62 FR 38652
July 18, 1997)
2006
(71 FR 61144
October 17, 2006)
Indicator
TSP
PM10
PM25
PM10
PM25
PM10
Ave. Time
24-hour
Annual
24-hour
Annual
24-hour
Annual
24-hour
Annual
24-hour
Annual
24-hour
Level
260 ug/m3
(primary)
150 ug/m3
(secondary)
75 ug/m3
(primary)
150 ug/m3
50 ug/m3
65 ug/m3
15 ug/m3
150 ug/m3
50 ug/m3
35 ug/m3
15 ug/m3
150 ug/m3
Form
Not to be exceeded more than once
per year
Annual average
Not to be exceeded more than once
per year on average over a 3 -year
period
Annual arithmetic mean, averaged
over 3 years
98th percentile, averaged over 3 years
Annual arithmetic mean, averaged
over 3 years8
Initially promulgated 99th percentile,
averaged over 3 years; when 1997
standards were vacated, the form of
1987 standards remained in place
(not to be exceeded more than once
per year on average over a 3 -year
period)
Annual arithmetic mean, averaged
over 3 years
98th percentile, averaged over 3 years
Annual arithmetic mean, averaged
over 3 years9
Not to be exceeded more than once
per year on average over a 3 -year
period
      When not specified, primary and secondary standards are identical.
     8 The level of the standard was to be compared to measurements made at sites that represent "community-wide air
     quality" recording the highest level, or, if specific constraints were met, measurements from multiple community-
     wide air quality monitoring sites could be averaged ("spatial averaging").
     9 The constraints on the spatial averaging criteria were tightened by further limiting the conditions under which some
     areas may average measurements from multiple community-oriented monitors to determine compliance (see 71 FR
     61165-61167).
     Draft Do Not Quote or Cite
1-6
March 2010

-------
 1          In April 1994, EPA announced its plans for the second periodic review of the criteria and
 2    NAAQS for PM, and promulgated significant revisions to the NAAQS in 1997 (62 FR 38652,
 3    judicial review, the revised standards were upheld in all respects. Natural Resources Defense
 4    Council v. EPA, 902 F. 2d 962 (D.C. Cir. 1990), cert, denied, 498 U.S. 1082 (1991). July 18,
 5    1997). Most significantly, EPA determined that although the PM NAAQS should continue to
 6    focus on particles less than or equal  to 10 jim in diameter, the fine and coarse fractions of PMio
 7    should be considered separately.  New standards were added, using PM2.s, referring to particles
 8    with a nominal median aerodynamic diameter less than or equal to 2.5 jim, as the indicator for
 9    fine particles. The PMio standards were retained for the purpose of regulating the coarse fraction
10    of PMio (referred to as thoracic coarse particles or coarse-fraction particles; generally including
11    particles with a nominal median aerodynamic diameter greater than 2.5 |im and less than or
12    equal to 10  jim, or PMio-2.s). The EPA established two new PM2.5 standards: an annual  standard
13    of 15 |ig/m3, based on the 3-year average of annual arithmetic mean PM2.5 concentrations from
14    single or multiple monitors sited to represent community-wide air quality; and a 24-hour
15    standard of 65 |ig/m3, based on the 3-year average of the  98th percentile of 24-hour PM2.s
16    concentrations at each population-oriented monitor within an area.  Also, EPA established a new
17    reference method for the measurement of PM2.5 in the ambient air and rules for determining
18    attainment of the new standards.  To continue to address thoracic coarse particles, the annual
19    PMio standard was  retained, while the form, but not the level, of the 24-hour PMio standard was
20    revised to be based on the 99th percentile of 24-hour PMio concentrations  at each monitor in an
21    area. The EPA revised the secondary standards by making them identical in all respects to the
22    primary standards.
23          Following promulgation of the revised PM NAAQS in 1997, petitions for review were
24    filed by a large number of parties, addressing a broad range of issues.  In May 1998, a three-
25    judge panel of the U.S. Court of Appeals for the District of Columbia Circuit issued an initial
26    decision that upheld EPA's decision to establish fine particle standards, holding that "the
27    growing empirical evidence demonstrating a relationship between fine particle pollution and
28    adverse health effects amply justifies establishment of new fine  particle standards." American
29    Trucking Associations v. EPA , 175  F. 3d 1027, 1055-56 (D.C. Cir. 1999) (rehearing granted in
30    part and denied in part, 195 F.  3d 4 (D.C. Cir. 1999), affirmed in part and reversed in part,
31    Whitman v.  American Trucking Associations., 531 U.S. 457 (2001). The panel also found "ample
32    support" for EPA's  decision to regulate coarse particle pollution, but vacated the 1997 PMio
33    standards, concluding in part that PMio is a "poorly matched indicator for coarse particulate
34    pollution" because it includes fine particles. Id. at 1053-55.  Pursuant to the court's decision,
35    EPA removed the vacated 1997 PMio standards from the  Code of Federal Regulations (CFR) (69
36    FR 45592, July 30, 2004) and deleted the regulatory  provision [at 40 CFR section 50.6(d)] that
      Draft Do Not Quote or Cite               1-7                            March 2010

-------
 1    controlled the transition from the pre-existing 1987 PMio standards to the 1997 PMio standards.
 2    The pre-existing 1987 PMio standards remained in place (65 FR 80776, December 22, 2000).
 3    The Court also upheld EPA's determination not to establish more stringent secondary standards
 4    for fine particles to address effects on visibility (175 F. 3d at 1027).
 5           More generally, the panel held (over a strong dissent) that EPA's approach to
 6    establishing the level of the standards in 1997, both for the PM and for the ozone (Os) NAAQS
 7    promulgated on the same day, effected "an unconstitutional delegation of legislative authority."
 8    Id. at 1034-40. Although the panel stated that "the factors EPA uses in determining the degree of
 9    public health concern associated with different levels of ozone and PM are reasonable," it
10    remanded the rule to EPA, stating that when EPA considers these factors for potential non-
11    threshold pollutants "what EPA lacks is any determinate criterion for drawing lines" to
12    determine where the standards should be set. Consistent with EPA's long-standing interpretation
13    and D.C.  Circuit precedent, the panel also reaffirmed its prior holdings that in setting NAAQS
14    EPA is "not permitted to consider the cost of implementing those standards" Id. at 1040-41.
15           On EPA's petition for rehearing, the panel adhered to its position on these points.
16    American Trucking Associations v. EPA, 195 F. 3d 4 (D.C. Cir.  1999). The full Court of
17    Appeals denied EPA's request for rehearing en bane, with five judges dissenting.  Id. at 13.  Both
18    sides filed cross appeals on these issues to the United States Supreme Court, which granted
19    certiorari. In February 2001, the Supreme Court issued a unanimous decision upholding EPA's
20    position on both the constitutional and cost issues. Whitman v.  American Trucking Associations.,
21    531 U.S.  457, 464, 475-76. On the constitutional issue, the Court held that the statutory
22    requirement that NAAQS be "requisite" to protect public health with an adequate margin of
23    safety  sufficiently cabined EPA's discretion, affirming EPA's approach of setting standards that
24    are neither more nor less stringent than necessary. The Supreme Court remanded the case to the
25    Court of Appeals for resolution of any remaining issues that had not been addressed  in that
26    court's earlier rulings. Id. at 475-76. In March 2002, the Court of Appeals rejected all
27    remaining challenges to the standards, holding under the traditional standard of review that
28    EPA's PM2.5 standards were reasonably supported by the administrative record and were not
29    "arbitrary and capricious." American Trucking Associations v. EPA, 283 F. 3d 355, 369-72 (D.C.
30    Cir. 2002).
31          In October 1997, EPA published its plans for the next periodic review of the  air quality
32    criteria and NAAQS for PM (62 FR 55201, October 23, 1997),  including the 1997 PM2.5
33    standards and the 1987 PMio standards. After CAS AC and public review of several drafts,
34    NCEA finalized the Air  Quality Criteria Document for Particulate Matter (henceforth, AQCD or
35    the "Criteria Document") in October 2004 (U.S. EPA, 2004) and OAQPS finalized an
3 6    assessment document, P articulate Matter Health Risk Assessment for Selected Urban Areas
      Draft Do Not Quote or Cite               1-8                            March 2010

-------
 1    (Abt, 2005), and a "Staff Paper," Review of the National Ambient Air Quality Standards for
 2    Particulate Matter:  Policy Assessment of Scientific and Technical Information, in December
 3    2005 (U.S. EPA, 2005). In conjunction with their review of the Staff Paper, CAS AC provided
 4    advice to the Adminstrator on revisions to the PM NAAQS (Henderson, 2005a). In particular,
 5    most CAS AC PM Panel members favored revising the level of the 24-hour PM2.5 primary
 6    standard in the range of 35 to 30 |ig/m3 with a 98th percentile form, in concert with  revising the
 7    level of the annual PM2.5 standard in the range of 14 to 13 |ig/m3 (Henderson, 2005a, p. 7).  For
 8    thoracic coarse particles, the Panel had reservations in recommending a 24-hour PMio-2.5 primary
 9    standard, and agreed that there was a need for more research on the health effects of thoracic
10    coarse particles (Henderson, 2005b).  With regard to secondary standards, most Panel members
11    strongly supported establishing a new, distinct secondary PM2 5 standard to protect urban
12    visibility (Henderson, 2005a, p. 9).
13           On January 17, 2006, EPA proposed to revise the primary and secondary NAAQS for PM
14    (71 FR 2620) and solicited comment on a broad range of options. Proposed revisions included:
15    revising the level of the 24-hour PM2.s primary standard to 35  |ig/m3; revising the form, but not
16    the level, of the annual  PM2.5 primary standard by tightening the constraints on the  use of spatial
17    averaging; replacing the 24-hour PMio primary standard with a 24-hour standard defined in
18    terms of a new indicator, PMio-2.5  10 set at a level of 70 |ig/m3; revoking the annual  PMio primary
19    standard; and revising the secondary standards by making them identical in all respects to the
20    proposed suite of primary standards for fine and coarse particles.11 Subsequent to the proposal,
21    CAS AC provided additional advice to EPA in a letter to the Administrator requesting
22    reconsideration of CASAC's recommendations for both the primary and secondary PM2 5
23    standards as well as the standards  for thoracic coarse particles (Henderson, 2006a).
24           On October 17,  2006, EPA promulgated revisions to the PM NAAQS to provide
25    increased protection of public health and welfare (71 FR 61144).  With regard to the primary and
26    secondary standards for fine particles, EPA revised the level of the 24-hour PM2.s standard to 35
27    ng/m3, retained the level of the annual PM2.5 standard at 15 |ig/m3, and  revised the form of the
28    annual PM2.5 standard by adding further constraints on the optional use of spatial averaging. The
29    EPA revised the secondary standards for fine particles by making them identical in all respects to
30    the primary standards.  With regard to the primary and secondary standards for thoracic coarse
      10 This proposed indicator was qualified so as to include any ambient mix of PM10.25 dominated by particles
      generated by high-density traffic on paved roads, industrial sources, and construction sources, and to exclude any
      ambient mix of particles dominated by rural windblown dust and soils and agricultural and mining sources.
      1J In recognition of an alternative view expressed by most members of the CAS AC PM Panel, the Agency also
      solicited comments on a subdaily (4 to 8 hour averaging time) secondary PM2 5 standard to address visibility
      impairment, within the range of 20 to 30 ug/m3 and with a form within the range of the 92nd to 98th percentile (71 FR
      2685).
      Draft Do Not Quote or Cite               1-9                            March 2010

-------
 1    particles, EPA retained the level and form of the 24-hour PMio standard (such that the standard
 2    remained at a level of 150 |ig/m3 with a one expected exceedance form), and revoked the annual
 3    PMio standard. The EPA also established a new Federal Reference Method (FRM) for the
 4    measurement of PMio-2.5 in the ambient air (71 FR 61212-13). Although the standards for
 5    thoracic coarse particles were not defined in terms of a PMio-2.5 indicator, the new FRM for
 6    PMio-2.5 was established to provide a basis for approving Federal Equivalent Methods (FEMs)
 7    and to promote gathering scientific data to support future reviews of the PM NAAQS.
 8          Following issuance of the final rule, CAS AC articulated its concern that "EPA's final
 9    rule on the NAAQS for PM does not reflect several important aspects of the CASAC's advice"
10    (Henderson et al, 2006b). With regard to the PM2.5 annual primary standard, CAS AC expressed
11    serious concerns regarding the decision to retain the level of the standard at 15 |ig/m3.  With
12    regard to EPA's final decision to retain the 24-hour PMio standard for thoracic coarse particles,
13    CAS AC acknowledged concerns associated with retaining this standard while recognizing the
14    need to have a standard in place to protect against effects associated with short-term exposures to
15    thoracic coarse particles.  With regard to EPA's final decision to revise the secondary PM2.s
16    standards to be identical in all respects to the revised primary PM2.5 standards, CAS AC
17    expressed concerns that CASAC's advice to establish a distinct secondary standard for fine
18    particles to address visibility impairment was not followed.

19    1.2.4  Litigation Related to the 2006 PM Standards
20          Several parties filed petitions for review following promulgation of the revised PM
21    NAAQS in 2006. These petitions addressed the following issues:  (1) selecting the level of the
22    primary annual PM2.s standard; (2) retaining PMio as the indicator of a standard for thoracic
23    coarse particles, retaining the level and form of the 24-hour PMio standard, and revoking the
24    PMio annual standard; and (3) setting the secondary PM2.s standards identical to the primary
25    standards.  On February 24, 2009, the  U.S. Court of Appeals for the District of Columbia Circuit
26    issued its opinion in the case American Farm Bureau Federation v. EPA, 559 F. 3d 512 (D.C.
27    Cir. 2009).  The court remanded the primary annual PM2.s NAAQS to EPA because EPA failed
28    to adequately explain why the standard provided the requisite protection from both short- and
29    long-term exposures to fine particles, including protection for at-risk populations. American
30    Farm Bureau Federation v. EPA, 559F.3d512, (D.C. Cir. 2009). With regard to the standards
31    for PMio, the court upheld EPA's decisions to retain the 24-hour PMio standard to provide
32    protection from thoracic coarse particle exposures and to revoke the annual PMi0. standard.
33    American Farm Bureau Federation at  533-38.  With regard to the secondary PM2.5 standards, the
34    court remanded the standards to EPA because the  Agency failed to adequately explain why

      Draft Do Not Quote or Cite               1-10                            March 2010

-------
 1    setting the secondary PM standards identical to the primary standards provided the required
 2    protection for public welfare, including protection from visibility impairment.
 3           The decisions of the court with regard to these three issues are discussed in chapters 2, 3
 4    and 4, respectively. The EPA is responding to the court's remands as part of the current review
 5    ofthePMNAAQS.

 6    1.2.5  Current PM NAAQS Review
 7          The EPA initiated the current review of the air quality criteria for PM in June 2007 with a
 8    general call for information (72 FR 35462, June 28, 2007).  In July 2007, EPA held two "kick-
 9    off' workshops on the primary and  secondary PM NAAQS, respectively (72 FR 34003 and
10    34005, June 20, 2007).12  These workshops provided an opportunity for the participants to
11    discuss the key policy-relevant issues around which EPA would structure this PM NAAQS
12    review and the most meaningful new science that would be available to inform our
13    understanding of these issues.
14          Based in part on the workshop discussions, EPA developed a draft IRP outlining the
15    schedule, process, and key policy-relevant questions that would guide the evaluation of the air
16    quality criteria for PM and the review of the primary and secondary PM NAAQS.  On November
17    30, 2007, EPA held a consultation with CASAC13 on the draft IRP (72 FR 63177, November 8,
18    2007), which included the opportunity for public comment.  The final IRP (US EPA, 2008a)
19    incorporated comments from CASAC and the public on the draft plan as well as input from
20    senior Agency managers.
21          As part of the process of preparing the PM ISA, NCEA hosted a peer review workshop in
22    June 2008 on preliminary drafts of key ISA chapters (73 FR 30391, May 27, 2008). The first
23    external review draft ISA (US EPA, 2008b) was reviewed by CASAC and the public at a
24    meeting held in April 2009 (74 FR 2688, February 19, 2009). Based on CASAC and public
25    comments, NCEA prepared a second draft ISA (US EPA, 2009b), which was reviewed by
26    CASAC and the public at a meeting held on October 5-6, 2009 (74 FR 46586, September 10,
27    2009).  Based on CASAC and public comments, NCEA prepared the final ISA (US EPA,
28    2009a; 74 FR 66353, December 15, 2009).
29          In preparing the REA documents that build on the scientific evidence presented in the
30    ISA, OAQPS released two planning documents: Particulate Matter National Ambient Air
      12 See workshop materials http://www.regulations.gov/search/Regs/home.htmMhome Docket ID numbers EPA-HQ-
      OAR-2007-0492-008; EPA-HQ-OAR-2007-0492-009; EPA-HQ-OAR-2007-0492-010; and EPA-HQ-OAR-2007-
      0492-012.
      13 The CASAC PM NAAQS Review Panel was consitituted to perform the statutorily required review of the criteria
      and standards for this review of the PM NAAQS. For more information on the CASAC PM Panel, see
      http://yosemite.epa.gov/sab/sabpeople.nsf/WebCommitteesSubcommittees/CASAC%20Particulate%20Matter%20R
      eview%20Panel.
      Draft Do Not Quote or Cite              1-11                            March 2010

-------
 1   Quality Standards: Scope and Methods Plan for Health Risk and Exposure Assessment and
 2   Particulate Matter National Ambient Air Quality Standards: Scope and Methods Plan for Urban
 3   Visibility Impact Assessment (henceforth, Scope and Methods Plans, US EPA, 2009c,d). These
 4   planning documents outlined the scope and approaches that staff planned to use in conducting
 5   quantitative assessments as well as key issues that would be addressed as part of the assessments.
 6   In designing and conducting the initial health risk and visibility impact assessments, we
 7   considered CASAC comments (Samet 2009a,b) on the Scope and Methods Plans made during an
 8   April 2009 consultation (74 FR 11580, March 18, 2009) as well as public comments..  Two draft
 9   assessment documents, Risk Assessment to Support the Review of the PM2.s Primary National
10   Ambient Air Quality Standards: External Review Draft - September 2009 (US EPA 2009e) and
11   Particulate Matter Urban-Focused Visibility Assessment - External Review Draft - September
12   2009 (US EPA, 2009f) were reviewed by CASAC and the public at a meeting held on October 5-
13   6, 2009.  Based on CASAC (Samet 2009c,d) and public comments, OAQPS staff revised these
14   draft documents and released second draft assessment documents (US EPA, 2010a,b) in January
15   and February 2010 (75 FR 4067, January 26, 2010) for CASAC and public review at an
16   upcoming meeting to be held on March 10-11, 2010.
17          A preliminary draft PA (US EPA, 2009g) was released in September 2009 for
18   informational purposes and to facilitate discussion with CASAC at the October 5-6, 2009
19   meeting on the overall structure, areas of focus,  and level of detail to be included in the PA.  This
20   first draft PA reflects consideration of CASAC's comments on the preliminary draft that
21   encouraged the development of a document focused on the key policy-relevant issues that draws
22   from and is not repetitive of information in the ISA and REAs.  This first draft PA draws from
23   the information presented in the final ISA and the two second draft assessment documents. We
24   plan to present an overview of this document at the upcoming CASAC meeting on March 10-11,
25   2010, and CASAC and public review of this document will occur during an upcoming
26   teleconference to be held on April 8-9, 2010 (75 FR 8062, February 23, 2010). We will consider
27   CASAC and public comments on this first draft PA and on the two draft REAs in preparing a
28   second draft PA, which will be released for CASAC and public review.

29   1.3    GENERAL APPROACH AND ORGANIZATION OF THIS DOCUMENT
30          This first draft PA includes staffs preliminary evaluation of the policy implications of
31   the scientific assessment of the evidence presented in the ISA and the results of quantitative
32   assessments based on that evidence presented in the second draft REAs.  Taken together, this
33   information informs preliminary staff conclusions and the identification of policy options for
34   consideration in addressing public health and welfare effects associated with exposure to ambient
35   PM.
     Draft Do Not Quote or Cite             1-12                           March 2010

-------
 1           Since the last review, much new information is now available on PM air quality and
 2    human health effects directly in terms of PM2.5 and, to a much more limited degree, PMio-2.5 and
 3    ultrafme particles (UFPs).  Since the purpose of this review is to evaluate the adequacy of the
 4    current standards, which separately address fine and thoracic coarse particles, staff is focusing
 5    this policy assessment and associated quantitative analyses primarily on the evidence related
 6    directly to PM2.5 and PMio-2.5-  In so doing, we are considering PMio-related evidence primarily
 7    to help inform our understanding of key issues and to help interpret and provide context for
 8    understanding the public health and welfare impacts of ambient fine and coarse particles. We are
 9    also considering the currently available evidence related to UFPs as well as PM2.5 components to
10    aid in considering whether or not there is support to consider standards with a different size
11    fraction and/or distinct standards focused on regulating specific PM2 5 components or categories
12    of fine particle sources.
13           Following this introductory chapter, this document is organized into two main parts:
14    review of the primary PM NAAQS (chapters 2 and  3) and review of the secondary PM NAAQS
15    (chapters 4 and  5).  Chapters 2 and 3 present staff observations and preliminary conclusions
16    related to review of the primary standards for fine and thoracic coarse particles, respectively.
17    Each chapter begins with a discussion of policy assessment approaches and focuses on both
18    evidence-based  and quantitative risk-based considerations.  Preliminary staff conclusions are
19    presented with regard to the adequacy of the current primary standards  and potential alternative
20    primary standards for consideration, in terms of indicators, averaging times, forms, and levels.
21    Chapter 4 focuses on PM-related visibility impairment, and presents staff observations and
22    preliminary conclusions with regard to the adequacy of the current standards and potential
23    distinct secondary standards for consideration, in terms of alternative indicators, averaging times,
24    forms, and levels. Chapter 5 focuses on other PM-related welfare effects, including effects on
25    climate, ecological  effects, and effects  on materials, and presents staff observations and
26    preliminary conclusions with regard to the adequacy of the current standards and the extent to
27    which information is available to support consideration of alternative standards.
      Draft Do Not Quote or Cite              1-13                            March 2010

-------
  1     1.4    REFERENCES

  2     Abt Associates (2005). Paniculate Matter Health Risk Assessment for Selected Urban Areas. Final Report.
  3            Bethesda, MD. Prepared for the Office of Air Quality Planning and Standards, U.S. Environmental
  4            Protection Agency, Contract No. 68-D-03-002.  EPA 452/R-05-007A. Available: 0

  5     Henderson, R. (2005a).  EPA's Review of the National Ambient Air Quality Standards for Paniculate Matter
  6            (Second Draft PM Staff Paper, January 2005): A review by the Paniculate Matter Review Panel of the EPA
  7            Clean Air Scientific Advisory Committee. June 6, 2005. Available:  http://www.epa.gov/sab/pdf/casac-05-
  8            007.pdf.

  9     Henderson, R. (2005b).  Clean Air Scientific Advisory Committee (CASAC) Review of the EPA Staff
10            Recommendations Concerning a Potential Thoracic Coarse PM Standard in the Review of the National
11            Ambient Air Quality Standards for Particulate Matter: Policy Assessment of Scientific and Technical
12            Information (Final PM OAQPS Staff Paper, EPA-452/R-05-005). September 15, 2005. Available:
13            http://www.epa.gov/sab/panels/casacpmpanel.html.

14     Henderson, R. (2006a). Letter from Dr. Rogene Henderson, Chair, Clean Air Scientific Advisory Committee to the
15            Honorable Stephen L. Johnson, Administrator, US EPA. Clean Air Scientific Advisory Committee
16            Recommendations Concerning the Proposed National Ambient Air Quality Standards for Particulate
17            Matter. March 21, 2006. Available: http://www.epa.gov/sab/pdf/casac-ltr-06-002.pdf

18     Henderson, R.; Cowling, E.; Crapo, J.D..; Miller, F. J.; Poirot, R.L.; Speizer, F.; Zielinska, B. (2006b).  Letter from
19            Clean Air Scientific Advisory Committee to the Honorable Stephen L. Johnson, Administrator, US EPA.
20            Clean Air Scientific Advisory Committee Recommendations Concerning the Final National Ambient  Air
21            Quality Standards for Particulate Matter.  September 29, 2006. Available:
22            http://www.epa.gov/sab/pdf/casac-ltr-06-002.pdf

23     Jackson (2009). Memo from Administrator Lisa P. Jackson to Elizabeth Craig, Acting Assistant Administrator for
24            OAR and Lek Kadeli, Acting Assistant Administrator for ORD.  Process for Reviewing the National
25            Ambient Air Quality Standards. May 21, 2009.  Available:
26            http://www.epa.gov/ttn/naaqs/pdfs/NAAQSReviewProcessMemo52109.pdf.

27     Samet, J. (2009a). Letter from Clean Air Scientific Advisory Committee to the Honorable Lisa P. Jackson,
28            Administrator, US EPA.  Consultation on EPA's Particulate Matter National Ambient Air Quality
29            Standards:  Scope and Methods Plan for Health Risk and Exposure Assessment.  May 21, 2009. Available:
30            http://YOsemite.epa. gov/sab/sabproduct.nsf/723FE644C5D758DF852575BD00763A32/$File/EPA-
31            CASAC-09-009-unsigned.pdf

32     Samet, J. (2009b). Letter from Clean Air Scientific Advisory Committee to the Honorable Lisa P. Jackson,
3 3            Administrator, US EPA.  Consultation on EPA's Particulate Matter National Ambient Air Quality
34            Standards:  Scope and Methods Plan for Urban Visibility Impact Assessment. May 21, 2009. Available:
35            http://yosemite.epa.gov/sab/sabproduct.nsf/OF63D7995F5850D5852575BD0077869C/$File/EPA-CASAC-
36            09-010-unsigned.pdf

37     Samet, J. (2009c). Letter from Clean Air Scientific Advisory Committee to the Honorable Lisa P. Jackson,
3 8            Administrator, US EPA.  Review of the Risk Assessment to Support the Review of the Particulate Matter
39            (PM) Primary National Ambient Air Quality Standards - External Review Draft (September 2009).
40            November 24, 2009.  Available:
41            http://vosemite.epa. gov/sab/sabproduct.nsf/BClECC5D539EF72385257678006D5754/$File/EPA-
42            CAS AC-10-003 -unsignedpdf

43     Samet, J. (2009d). Letter from Clean Air Scientific Advisory Committee to the Honorable Lisa P. Jackson,
44            Administrator, US EPA.  Review of Particulate Matter Urban-Focused Visibility Assessment (External

       Draft Do Not Quote or Cite                1-14                                 March  2010

-------
  1            Review Draft, September 2009). November 24, 2009. Available:
  2            http://vosemite.epa.gov/sab/sabproduct.nsf/15872217938041F685257678006A26E3/$File/EPA-CASAC-
  3            10-002-unsigned.pdf

  4    US Department of Health, Education and Welfare (DHEW). (1969). Air Quality Criteria for Paniculate Matter.
  5            U.S. Government Printing Office, Washington DC, AP-49.

  6    US EPA (2004). Air Quality Criteria for Paniculate Matter. National Center for Environmental Assessment, Office
  7            of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711;
  8            report no. EPA/600/P-99/002aF and EPA/600/P-99/002bF. October 2004. Available:
  9            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm cr cd.html

10    US EPA (2005). Review of the National Ambient Air Quality Standards for Paniculate Matter: Policy Assessment
11            of Scientific and Technical Information, O AQPS Staff Paper.  Research Triangle Park, NC 27711: Office
12            of Air Quality Planning and Standards; report no. EPA EPA-452/R-05-005a.  December 2005. Available:
13            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm cr sp.html

14    US EPA (2008a).  Integrated Review Plan for the National Ambient Air Quality Standards for Paniculate Matter.
15            National Center for Environmental Assessment and Office of Air Quality Planning and Standards, U.S.
16            Environmental Protection Agency, Research Triangle Park, NC. Report No.  EPA 452/R-08-004. March
17            2008. Available at: http://www.epa.gov/ttn/naaqs/standards/pm/s_pm 2007_pd.html.

18    US EPA (2008b). Integrated Science Assessment for Paniculate Matter: First External Review Draft. National
19            Center for Environmental Assessment-RTF Division, Office of Air Quality Planning and Standards,
20            Research Triangle Park, NC. EPA/600/R-08/139 and 139A.  December 2008. Available:
21            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_2007  isa.html.

22    US EPA (2009a).  Integrated Science Assessment for Paniculate Matter: Final Report. National Center for
23            Environmental Assessment-RTF Division, Office of Air Quality Planning and Standards, Research
24            Triangle Park, NC. EPA/600/R-08/139F. December 2009. Available:
25            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_2007  isa.html.

26    US EPA (2009b). Integrated Science Assessment for Paniculate Matter: Second External Review Draft. National
27            Center for Environmental Assessment-RTF Division, Office of Air Quality Planning and Standards,
28            Research Triangle Park, NC. EPA/600/R-08/139B.  July 2009. Available:
29            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_2007  isa.html.

30    US EPA (2009c).  Paniculate Matter National Ambient Air Quality Standards:   Scope and Methods Plan for Health
31            Risk and Exposure Assessment. Office of Air Quality Planning and Standards, U.S. Environmental
32            Protection Agency, Research Triangle Park, NC. EPA-452/P-09-002.  February 2009.  Available:
33            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_2007_pd.html.

34    US EPA (2009d). Paniculate Matter National Ambient Air Quality Standards:   Scope and Methods Plan for Urban
3 5            Visibility Impact Assessment. Office of Air Quality Planning and Standards, U.S. Environmental
36            Protection Agency, Research Triangle Park, NC. EPA-452/P-09-001. February 2009. Available:
37            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm 2007_pd.html.

38    US EPA (2009e).  Risk Assessment to Support the Review of the PM Primary National Ambient Air Quality
39            Standards - External Review Draft. Office of Air Quality Planning and Standards, U.S. Environmental
40            Protection Agency, Research Triangle Park, NC. EPA-452/P-09-006.  September 2009. Available:
41            http://www.epa.gov/ttn/naaqs/standards/pm/sjm 2007  risk.html.

42    US EPA (20091). Paniculate Matter Urban-Focused Visibility Assessment - External Review Draft.  Office of Air
43            Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle  Park, NC.


      Draft Do Not Quote or Cite                1-15                                March 2010

-------
 1            EPA-452/P-09-005.  September 2009. Available:
 2            http://www.epa.gov/ttn/naaqs/standards/pm/sjm  2007 risk.html.

 3    US EPA (2009g).  Policy Assessment for the Review of the Particulate Matter National Ambient Air Quality
 4            Standards - Preliminary Draft. Office of Air Quality Planning and Standards, U.S. Environmental
 5            Protection Agency, Research Triangle Park, NC. EPA-452/P-09-007. September 2009. Available:
 6            http://www.epa.gov/ttn/naaqs/standards/pm/sjm  2007ja.html.

 7    US EPA (2010a).  Quantitative Risk Assessment for Particulate Matter - Second External Review Draft. Office of
 8            Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC.
 9            EPA-452/P-10-001.  February 2010.  Available:
10            http://www.epa.gov/ttn/naaqs/standards/pm/sjm  2007 risk.html.

11    US EPA (2010b).  Particulate Matter Urban-Focused Visibility Assessment - Second External Review Draft. Office
12            of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park,
13            NC. EPA-452/P-10-002. January 2010. Available:
14            http://www.epa.gOv/ttn/naaqs/standards/pm/s  pm  2007 risk.html.
      Draft Do Not Quote or Cite                 1-16                                March 2010

-------
 1    2   REVIEW OF THE PRIMARY STANDARDS FOR FINE PARTICLES

 2           This chapter presents preliminary staff conclusions with regard to the adequacy of the
 3    current suite of primary PM2.5 standards and the alternative primary standards for fine particles
 4    that are appropriate for consideration in this review. Our assessment of these issues is framed by
 5    a series of key policy-relevant questions, which expand upon those presented in the Integrated
 6    Review Plan (IRP) (US EPA, 2008a) at the outset of this review.  The answers to these questions
 7    will inform decisions on whether, and if so how, to revise the current suite of primary fine
 8    particle standards.
 9           Staff notes that final decisions regarding the primary standards must draw upon scientific
10    information and analyses about health effects and risks, as well as judgments about how to deal
11    with the range of uncertainties that are inherent in the scientific evidence and analyses.
12    Ultimately, the final decisions are largely public health policy judgments.  Our approach to
13    informing these judgments, discussed more fully below, recognizes that the available health
14    effects evidence generally reflects a continuum consisting of ambient levels at which scientists
15    generally agree that health effects are likely to occur through lower levels at which the likelihood
16    and magnitude of the response become increasingly uncertain
17           Our current approach for reviewing the primary standards for fine particles is presented
18    in section 2.1.  Our preliminary conclusions regarding the adequacy of the current suite of
19    primary PM2.5 standards are presented in section 2.2, focusing on both evidence-based and
20    quantitative risk-based considerations.  Section 2.3 presents our initial conclusions with respect
21    to alternative fine particle standards, focusing on each of the basic elements of the standards:
22    pollutant indicator (section 2.3.1), averaging time (section 2.3.2), form (section 2.3.3), and level.
23    We have evaluated separately the protection that a suite of PM2.5 standards with alternative levels
24    would likely provide against effects associated with long-term exposures (section 2.3.4) and
25    short-term exposures (section 2.3.5). These separate evaluations provide the basis for
26    preliminary integrated conclusions on alternative suites of standards that would appropriately
27    protect against effects associated with both long- and short-term exposures to fine particles
28    (section 2.3.6)  Section 2.4 summarizes all preliminary staff conclusions on the primary fine
29    particle standards. The next draft of this chapter will conclude with an initial overview of key
30    uncertainties and suggested future research areas and data collection efforts (section 2.5).

31    2.1    APPROACH
32           Staffs approach for reviewing the current primary PM2.5 standards builds upon and
33    broadens the approaches used in previous PM NAAQS reviews. Our current approach is based
34    on the updated scientific and technical information presented in the Integrated Science
            Draft Do Not Quote or Cite            2-1                         March 2010

-------
 1    Assessment (ISA) and second draft quantitative risk assessment (RA).  These assessments take
 2    into consideration the currently available scientific information as well as enhanced tools and
 3    methods for informing the current review.
 4          The past and current approaches described below are all based most fundamentally on
 5    using information from epidemiological studies to inform the selection of PM standards that, in
 6    the Administrator's judgment, protect public health with an adequate margin of safety. Such
 7    information can be in the form of air quality distributions over which health effect associations
 8    have been observed, or in the form of concentration-response functions that support quantitative
 9    risk assessment. However, evidence- and risk-based approaches using information from
10    epidemiological studies to inform decisions on PM standards are complicated by the recognition
11    that no population threshold, below which it can be concluded with confidence that PM-related
12    effects do not occur, can be discerned from the available evidence. As a result, any approach to
13    reaching decisions on what standards are appropriate necessarily requires judgments about how
14    to translate the information available from the epidemiological studies into a basis for
15    appropriate standards, which includes consideration of how to weigh the uncertainties in reported
16    associations across the distributions of PM concentrations in the studies or in quantitative
17    estimates of risk.  Such approaches are  consistent with setting standards that are neither more nor
18    less stringent than necessary, recognizing that a zero-risk standard is not required by the CAA.

19    2.1.1  Approaches Used in Previous  Reviews
20          Staff has considered policy assessment approaches used in past reviews to inform the
21    approach we are using in this review to reach preliminary conclusions regarding the adequacy of
22    the current standard and alternative standards that  are appropriate to consider in this review. We
23    begin this section with a review of the approach used to set the original fine particle standards in
24    1997 (section 2.1.1.1). The  approach used to review and ultimately support revisions to these
25    standards in 2006 is discussed in section 2.1.1.2. Litigation related to the standards finalized in
26    2006, including the remand of the primary PM2.5 annual standard are then discussed in section
27    2.1.1.3.

28          2.1.1.1 Review Completed in 1997
29          In setting the 1997 primary PM2.5 annual and 24-hour standards, the Agency relied
30    primarily on an evidence-based approach that focused on epidemiological evidence, especially
31    from short-term exposure studies of fine particles judged to be the strongest evidence at that
32    time. The EPA did not place much weight on quantitative risk estimates from the very limited
33    risk assessment conducted, but did conclude that the assessment results confirmed the general
34    conclusions drawn from the epidemiological evidence that a serious public health problem was

            Draft Do Not Quote or Cite           2-2                         March 2010

-------
 1    associated with ambient PM levels allowed under the then current PMio standards (62 FR
 2    38665/1, July 18, 1997).
 3            The EPA considered the epidemiological evidence and data on air quality relationships
 4    to set an annual PM2.5 standard that was intended to be the "generally controlling" standard; i.e.,
 5    the primary means of lowering both long- and short-term ambient concentrations of PM2.5.1 In
 6    conjunction with the annual standard, EPA also established a 24-hour PM2.5 standard to provide
 7    supplemental protection against days with high peak concentrations, localized "hotspots," and
 8    risks arising from seasonal emissions that might  not be well controlled by a national annual
 9    standard. (62 FR 38669/3). Recognizing that there are various ways to combine two standards to
10    achieve an  appropriate degree of public health protection, such as an approach that only
11    considered short- and long-term exposure evidence, analyses, and standards independently, EPA
12    concluded that the selected approach based on a  generally controlling annual  standard was the
13    most effective and efficient approach.  This conclusion was based in part on one of the key
14    observations from the quantitative risk assessment, that much if not most of the aggregate annual
15    risk associated with short-term exposures results from the large number of days during which the
16    24-hour average concentrations are in the low- to mid-range, below the peak 24-hour
17    concentrations. As a result, lowering a wide range of ambient 24-hour PM2.5  concentrations by
18    means of a generally controlling annual standard, as opposed to focusing on control of peak 24-
19    hour concentrations, was determined to be the most effective and efficient way to reduce total
20    population  risk (62 FR 38670 to 38671).
21           In setting the level of the annual standard in 1997, EPA first determined a level for the
22    annual standard based on the short-term exposure studies, and then considered whether the key
23    long-term exposure studies suggested the need for a lower level.  While recognizing that health
24    effects may occur over the full range of concentrations observed in the studies, EPA concluded
25    that the strongest evidence for short-term PM2.5-related effects occurs at concentrations near the
26    long-term (e.g., annual) average in the short-term exposure studies.  Given the serious nature  of
27    the potential effects, EPA selected a level for the annual standard at or below the long-term mean
28    concentrations in studies providing evidence of associations with short-term exposures, placing
29    greatest weight on those short-term exposure studies that reported clearly statistically significant
30    associations with mortality and morbidity effects (62 FR 38676/1).  Further consideration of the
      1 In so doing, EPA noted that an annual standard would focus control programs on annual average PM2 5
      concentrations, which would generally control the overall distribution of 24-hour exposure levels, as well as long-
      term exposure levels, and would also result in fewer and lower 24-hour peak concentrations. Alternatively, a 24-
      hour standard that focused controls on peak concentrations could also result in lower annual average concentrations.
      Thus, EPA recognized that either standard could provide some degree of protection from both short- and long-term
      exposures, with the other standard serving to address situations where the daily peaks and annual averages are not
      consistently correlated (62 FR 38669).
             Draft Do Not Quote or Cite           2-3                          March 2010

-------
 1    average PM2.5 concentrations across the cities in the key long-term exposure studies of mortality
 2    and respiratory effects in children did not provide a basis for establishing a lower annual
 3    standard level. Because the annual standard level selected was below the range of annual
 4    concentrations most strongly associated with both short- and long-term exposure effects, and
 5    because even small changes in annual means in this concentration range could make significant
 6    difference in overall risk reduction and total population exposures, EPA concluded that this
 7    standard would provide an adequate margin of safety against effects observed in these
 8    epidemiological studies (62 FR 68676/3).
 9           The selection of the level of the annual standard was done in conjunction with having
10    first selected the form of the annual standard to be based on the concentration measured at a
11    single monitor sited to represent community-wide air quality, or a value resulting from an
12    average of measurements from multiple community-wide air quality monitoring sites that met
13    specific criteria and constraints ("spatial averaging"). This decision emphasized consistency
14    with the types of air quality measurements that were used in the relevant epidemiological studies.
15    In reaching this decision, EPA recognized the importance of ensuring that spatial averaging
16    would not result in inequities in the level of protection provided by the PM2.5 standards in some
17    areas. Because the annual  standard, defined in terms of single or averaged community-wide air
18    quality monitoring sites, could not be expected to offer an adequate margin of safety against the
19    effects of all potential short-term exposures in areas with strong local  or seasonal sources that
20    could not be directly evaluated in the epidemiological studies, EPA set the level of the 24-hour
21    standard to supplement the control afforded by the  annual standard based on air quality
22    relationships between annual and 24-hour concentrations. This approach was intended to
23    provide an adequate margin of safety against infrequent or isolated peak concentrations that
24    could occur in areas that attain the annual standard  (62 FR 38677).

25           2.1.1.2 Review Completed in 2006
26           In 2006, EPA used  a different evidence-based approach to assess the appropriateness of
27    the levels of the 24-hour and annual PM2.5 standards.  Based on an expanded body of
28    epidemiological evidence that was stronger and more robust, including both short- and long-term
29    exposure studies, the Administrator decided that using evidence of effects associated with
30    periods of exposure that were most closely matched to the averaging time of each standard was
31    the most appropriate public health  policy approach for evaluating the  scientific evidence to
32    inform selecting the level of each standard.  Thus, the Administrator relied upon evidence from
33    the short-term exposure studies as the principal basis for selecting the level of the 24-hour PM2.5
34    standard, with the 24-hour  standard designed to protect against effects associated with short-term
35    exposures.  The Administrator relied upon evidence from long-term exposure studies as the
             Draft Do Not Quote or Cite           2-4                         March 2010

-------
 1    principal basis for selecting the level of the annual PM2.5 standard, with the annual standard
 2    designed to protect against effects associated with long-term exposures.
 3           With respect to quantitative risk-based considerations, the Administrator recognized that
 4    the analyses conducted for this review were based on "a more extensive body of data and [was]
 5    more comprehensive in scope than the assessment conducted in the last review, but was mindful
 6    that significant uncertainties continue[d] to underlie the resulting risk estimates" (71 FR 61168/2,
 7    October 17, 2006).2 The Administrator determined that the estimates of risks likely to remain
 8    upon attainment of the 1997 suite of PM2.5 standards were indicative of risks that could be
 9    reasonably judged important from a public health perspective, and, thus, supported revision of
10    the standards.  However, the Administrator judged that the quantitative risk assessment had
11    important limitations and did not provide an appropriate basis for selecting either the level of the
12    24-hour or annual PM2.5 standard (71 FR 61174/1-2). The Administrator more heavily weighed
13    the implications of the uncertainties associated with the quantitative risk assessment than the
14    Clean Air Scientific Advisory Committee (CASAC) apparently did in their comments on the
15    proposed rulemaking, where CASAC stated,  "[w]hile the risk assessment is subject to
16    uncertainties, most of the PM Panel found EPA's risk assessment to be of sufficient quality to
17    inform its recommendations... The risk analyses indicated that the uncertainties would increase
18    rapidly below an annual level  of 13 |ig/m3 - and that was the basis for the PM Panel's
19    recommendation of 13 |ig/m3 as the lower bound for the annual PM2 5 standard level"
20    (Henderson, 2006a, p.3).
21           With regard to final decisions on the primary annual PM2.5 standard, the Administrator
22    placed the greatest weight for  determining the level of the annual standard on the long-term
23    means of the concentrations associated with mortality effects in the two key long-term exposure
24    studies in the record, the American  Cancer Society (ACS) and Harvard Six Cities studies (71 FR
25    at 61172 to 61177). Important validation and reanalyses  of the original ACS and Harvard Six
26    Cities studies provided "evidence of generally robust associations  and provide[d] a basis for
27    greater confidence in the reported associations than in the last review, for example, in the extent
28    to which they have made progress in understanding the importance of issues related to co-
29    pollutant confounding and the specification of statistical models."   Furthermore, the extended
30    ACS study provided "new evidence of mortality related to lung cancer and further
       Uncertainties identified in the quantitative risk assessment were generally related to a lack of clear understanding
      of a number of important factors, including, for example, the shape of concentration-response functions, particularly
      when, as here, effect thresholds could neither be discerned nor determined not to exist; issues related to selection of
      appropriate statistical models for the analysis of the epidemiologic data; the role of potentially confounding and
      modifying factors in the concentration-response relationships; issues related to simulating how PM2 5 air quality
      distributions would likely change in any given area upon attaining a particular standard, since strategies to reduce
      emissions were not yet defined; and whether there would be differential reductions in the many components within
      PM2 5 and, if so, whether this would result in differential reductions in risk (71 FR 61168/2).
             Draft Do Not Quote or Cite           2-5                         March 2010

-------
 1    substantiate^] the statistically significant associations with cardiorespiratory-related mortality
 2    observed in the original studies" (71 FR 61172/1-2).  The Administrator also recognized the
 3    availability of long-term exposure studies that provided evidence of respiratory morbidity,
 4    including changes in lung function measurements and decreased growth in lung function as
 5    reported in the 24-Cities study and the Southern California Children's Health Study,
 6    respectively.  In retaining the level of the annual standard at 15 |ig/m3, the Administrator selected
 7    a level that was "appreciably below" the long-term average concentrations reported in the key
 8    long-term mortality studies and "basically at the same level" as the long-term average
 9    concentrations in the key long-term respiratory  morbidity studies. In the judgment of the
10    Administrator, the two long-term respiratory morbidity studies provided an uncertain basis for
11    setting the level of a national standard, and therefore, did "not warrant setting a lower level for
12    the annual standard than the level warranted based on the key mortality studies" (71 FR
13    61176/3).
14           In considering the form of the primary annual PM2.5 standard, the Administrator retained
15    the form of the standard as an annual arithmetic mean, averaged over 3 years with modifications
16    that strengthen the standard by tightening the criteria for use of spatial averaging.  Specifically,
17    the Administrator narrowed the circumstances under which spatial averaging may be utilized.
18    Analyses conducted in the review completed in 2006, based on a much larger set of PM2.5 air
19    quality data than was available for the review completed in 1997, provided  evidence concerning
20    the potential for disproportionate impacts on potentially vulnerable subpopulations. Specifically,
21    the results of the analyses suggested that "the highest concentrations in an area tend to be
22    measured at monitors located in areas where the surrounding population [was] more likely to
23    have lower education and income levels, and higher percentages of minority populations" (71 FR
24    61166/2, see also US EPA, 2005, section 5.3.6.1; Schmidt et al., 2005, Attachment A/Analysis
25    7).3
26           In deciding to revise the level  of the 24-hour PM2.s standard from 65 |ig/m3 to 35 |ig/m3,
27    the Administrator placed the greatest weight on the much expanded body of evidence from short-
28    term exposure studies,  with a focus on U.S. and Canadian  studies that had been reanalyzed, as
29    appropriate, to address statistical modeling issues and that used relatively reliable air quality
30    data.  A comprehensive evaluation considered and weighed a variety of evidence, including
      3 As summarized in footnote 29 at 71 FR 61166/2, the 2004 AQCD noted that some epidemiologic studies, most
      notably the ACS study of associations between long-term PM2 5 exposure and mortality, reported larger effect
      estimates in the cohort subgroup with lower education levels (US EPA, 2004, p 8-103). The 2004 AQCD also noted
      that lower education level may be a marker for lower socioeconomic status (SES) that may be related to increased
      vulnerability to the effects of fine particle exposures, for example, as a result of greater exposure from proximity to
      sources such as roadways and industry, as well as other factors such as poorer health status and access to health care
      (US EPA, 2004, section 9.2.4.5).
             Draft Do Not Quote or Cite            2-6                         March 2010

-------
 1    biological plausibility of associations between various pollutants and health outcomes, and
 2    focused on the stability of the size of the effect estimates in time-series studies using both single-
 3    and multi-pollutant models, rather than just looking at statistical significance in a large number
 4    of alternative models to simplistically delineate between real and suspect associations (71 FR
 5    61170/2). Emphasis was placed on those studies that provided evidence of statistically
 6    significant associations in areas that would have met the then current annual and 24-hour PM2.5
 7    standards during the time of the study. The Administrator recognized that these studies provided
 8    no evidence of clear effect thresholds or lowest-observed effect levels.  Nonetheless, in focusing
 9    on the 98th percentile air quality values in these studies, the Administrator sought to establish a
10    standard level that would require improvements in air quality generally in areas in which the
11    distribution of daily short-term exposure to PM2 5 could reasonably be expected to be associated
12    with serious health effects. The Administrator recognized that although future air quality
13    improvement strategies in any particular area were not yet defined, most such strategies were
14    likely to move a broad distribution of PM2.5 air quality values in an area lower, resulting in
15    reductions in risk associated with exposures to PM2.5 levels across a wide range of concentrations
16    and not just at the 98th percentile concentrations (71 FR 61168/3).

17    2.1.2   Remand of Primary Annual PM2.5 Standard
18           As noted above in sectionl.2.4, several parties filed petitions for review following
19    promulgation of the revised PM NAAQS in 2006. These petitions challenged several aspects of
20    the final rule including the selection of the level of the primary PM2 5 annual standard.  More
21    specifically, petitioners representing environmental groups (American Lung Association,
22    Environmental Defense, and the National Parks Conservation Association) and several states and
23    state agencies argued that the decision to retain the level of the annual PM2 5 standard at 15
24    |ig/m3 was "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with
25    law."  42 U.S.C. 7607(d)(9).  The primary 24-hour PM2.5 standard was not challenged by any of
26    the litigants and, thus, not considered in the court's review and final decision.
27           On judicial review, the D.C. Circuit remanded the primary annual PM25 NAAQS to EPA
28    because the Agency failed to adequately explain why the annual standard provided the requisite
29    protection from both short- and long-term exposures to fine particles including protection for
30    susceptible populations. American Farm Bureau Federation v. EPA, 559 F. 3d 512 (D.C. Cir.
31    2009).  With respect to human health protection from short-term PM2 5 exposures, the court
32    considered the different approaches used by EPA in the 1997 and 2006 PM NAAQS decisions as
33    summarized above.  The court found that EPA failed to adequately explain why a 24-hour PM2 5
34    standard by itself would provide the protection needed from short-term exposures and remanded
35    the annual PM2.5 standard to EPA "for further consideration of whether it is set at a level

             Draft Do Not Quote or Cite            2-7                        March 2010

-------
 1    requisite to protect the public health while providing an adequate margin of safety from the risk
 2    of short-term exposures to PM2.5." American Farm Bureau Federation, 559 F. 3d at 520-24.
 3          With respect to protection from long-term exposure to fine particles, the court found that
 4    EPA failed to adequately explain how the current primary annual PM2.5 standard provided an
 5    adequate margin of safety for children and other susceptible populations.  The court found that
 6    EPA did not provide a reasonable explanation of why certain morbidity studies, including a
 7    study of children in Southern California showing lung damage associated with long-term PM2.5
 8    exposure (Gauderman et.al, 2000) and a multi-city study (24-Cities Study) evaluating decreased
 9    lung function in children associated with long-term PM2.5 exposures (Raizenne et al., 1996), did
10    not call for a more stringent annual PM2 5 standard.  Id. at 522-23. Specifically, the court found
11    that:
12
13          EPA was unreasonably confident that, even though it relied solely upon long-term
14          mortality studies, the revised standard would provide an adequate margin of safety with
15          respect to morbidity among children. Notably absent from the final rule, moreover, is
16          any indication of how the standard will adequately reduce risk to the elderly or to those
17          with certain heart or lung diseases despite (a) the EPA's determination in its proposed
18          rule that those subpopulations are at greater risk from exposure to fine particles and (b)
19          the evidence in the record supporting that determination. Id. at 525.
20
21          Petitioners also objected to the EPA analysis of the long-term exposure studies arguing
22    that the EPA "unreasonably focused" upon the long-term mean ambient concentrations of PM2.5
23    in the ACS (17.7 |ig/m3) and the Harvard Six Cities Study (18  |ig/m3) and then set a level below
24    those concentrations for the annual PM2 5 standard.  Specifically, these petitioners claimed that
25    "this approach violates the requirement of the Clean Air Act (CAA) that the NAAQS provide 'an
26    adequate margin of safety' which requires that EPA 'err on the side of caution'" and argued that
27    the level of the annual standard should be revised lower because the most recent data from these
28    studies "showed adverse health effects in years when the mean ambient concentration of PM2.5
29    was below 15 jig/m3." Id. at 526. The court rejected these arguments stating:  "The EPA,
30    mindful of its  obligation to set a  standard 'not lower or higher than is necessary.. .to protect
31    public health,' Whitman, 531  U.S. at 475-76, reasonably decided to address long-term mean
32    concentrations in the ACS and Six Cities studies... We approved a similar approach to assuring
33    an adequate margin of safety in ATAIII and  we do so again here." Id. at 527.
34          Petitioners also challenged the Agency's decision not to rely upon the quantitative risk
35    assessment in making final decisions on the level of the primary annual PM2.s standard.  More
36    specifically, these  petitioners argued that "[b]oth CASAC and EPA staff concluded that the
37    Agency's risk assessment... was of sufficient quality to deserve consideration in determining
38    the level of the annual standard necessary to provide an adequate margin of safety" and that, in
            Draft Do Not Quote or Cite           2-8                        March 2010

-------
 1    retaining the level of the annual standard at 15 |ig/m3 significant public health impacts remained
 2    (State brief, pp. 22 to 23). The court rejected arguments that EPA was compelled to use a risk-
 3    based approach to determine the level of the standards, or otherwise compelled to use the risk
 4    assessment quantitatively in the standard-setting process, concluding that EPA "reasonably
 5    analyzed the risk assessment," and deferred to the EPA's assessment of "scientific data within its
 6    technical expertise." Id at 529-530.
 7          In remanding the primary annual standard for reconsideration, the court did not vacate the
 8    annual standard. Id. at 530.

 9    2.1.3  Current Approach
10          Staffs approach in this review is founded on a much expanded body of epidemiological
11    evidence, more extensive air quality data and analyses, and a more comprehensive risk
12    assessment relative to the information available in past reviews.  As a result, our approach to
13    reaching conclusions about the adequacy of the current suite of PM2.5 standards and potential
14    alternative standards that are appropriate for consideration is broader and more integrative than
15    in past reviews. Our approach takes into account both evidence-based and risk-based
16    considerations, and the uncertainties related to both types  of information, to inform the
17    preliminary conclusions presented in this first draft Policy Assessment (PA). In so doing, we are
18    seeking to provide as broad an array of options as is supportable by the available information,
19    recognizing that the selection of a specific approach to reaching final decisions on the primary
20    PM2 5 standards will reflect the judgments  of the Administrator as to what weight to place on the
21    various approaches and types of information presented in the final PA.
22          As an initial matter, we believe it is most  appropriate to consider the protection against
23    PM2 s-related mortality and morbidity effects, associated with both long- and short-term
24    exposures, afforded by the annual and 24-hour standards taken together, rather than to consider
25    each standard separately. In so doing, we look at the  types of evidence that can inform each
26    standard, then integrate the results of those considerations to reach preliminary conclusions about
27    the current and alternative suites of standards. This approach reflects the recognition that
28    changes in PM2 5 air quality designed to meet an annual standard would likely result not only in
29    lower annual average concentrations but also in fewer and lower peak 24-hour concentrations.
30    Conversely, we recognize that changes designed to meet a 24-hour standard would result not
31    only in fewer and lower peak 24-hour concentrations  but also in lower annual average
32    concentrations, especially to the  extent that changes of a more regional, rather than local, nature
33    occur. The extent to which these two standards are interrelated in any given area depends in
34    large part on the relative levels of the standards, the peak-to-mean ratios that characterize air
            Draft Do Not Quote or Cite            2-9                         March 2010

-------
 1    quality patterns in an area, and whether changes in air quality designed to meet a given suite of
 2    standards is of a more regional or more localized nature.
 3           Our consideration of the protection afforded by the current and alternative suites of
 4    standards focuses first on PM2.5-related health effects associated with long-term exposures, for
 5    which quantitative estimates of risks to public health are appreciably larger; we then also focus
 6    on effects associated with short-term exposures. In both cases, we place greatest weight on
 7    associations that have been judged in the ISA to be causal and likely causal, while also
 8    considering associations judged to be suggestive of a causal relationship or that focus on specific
 9    susceptible populations. We focus on studies conducted in the U.S.  and Canada and place
10    relatively greater weight on statistically significant associations that yield relatively more precise
11    effect estimates and that are judged to be robust to confounding by other air pollutants.  In the
12    case of short-term exposure studies, we place greatest weight on large multi-city studies, while
13    also considering associations in  single-city studies.
14           As part of our evidence-based approach, we first evaluate the evidence from long-term
15    exposure studies, as well as the uncertainties and limitations in that evidence, to assess the
16    degree  to which the current and  alternative suites of standards can be expected to protect against
17    effects  related to long-term exposures.  As in past reviews, we believe this evidence can most
18    directly help inform consideration of the protection afforded by an annual PM2.5 standard. In
19    evaluating the long-term exposure evidence, we look at the aggregate long-term mean PM2 5
20    concentration in each study, the  range of long-term mean PM2.5 concentrations across cities, and
21    the distribution of city-specific means in terms of the standard deviation or interquartile range, to
22    the extent such data are available.  In considering the adequacy of the current suite of PM2 5
23    standards in protecting against long-term exposure-related effects, we consider whether
24    associations from long-term exposure studies have been reported across areas in which the
25    aggregate long-term study mean concentrations are at or below the level of the current annual
26    standard. We conclude that such long-term exposure studies can reasonably be viewed as calling
27    into question the adequacy of the current annual standard.
28           In considering what alternative standards would be protective against effects observed in
29    such long-term exposure studies, we first note the absence of any discernable threshold within
30    the range of long-term mean concentrations reported in the long-term exposure studies. While
31    recognizing that health effects may occur over the full range of concentrations observed in the
32    studies, we believe that it is reasonable to conclude that the evidence of association is strongest
33    down to somewhat below the aggregate mean concentration, such as down to one standard
34    deviation below the mean or to the lower end of the interquartile range, which includes the range
35    in which the data in the study are most concentrated. We also believe it is appropriate to
36    consider the long-term mean concentration at the point where the confidence interval becomes
             Draft Do Not Quote or Cite           2-10                        March 2010

-------
 1    notably wider, suggestive of a concentration below which the association becomes appreciably
 2    more uncertain and the possibility that an effects threshold may exist becomes more likely.
 3    Based on these considerations, we identify a range of alternative annual standard levels that we
 4    judge to be appropriate to consider for protecting against PM2.5-related long-term exposure
 5    effects based on the evidence available in this review.
 6           We next evaluate the evidence from short-term exposure studies, as well as the
 7    uncertainties and limitations in that evidence, to assess the degree to which the current and
 8    alternative suites of standards can be expected to protect against effects related to short-term
 9    exposures. As  in past reviews, it is  staffs view that this evidence can help inform consideration
10    of the protection afforded by both an annual and 24-hour standard.4  In evaluating the short-term
11    exposure evidence, we look both at the aggregate long-term mean PM2 5 concentrations in key
12    studies as well as the distributions of 24-hour PM2.5 concentrations, with a focus on the 98th
13    percentile concentrations to match the form of the current 24-hour standard, to the extent such
14    data are available.  In evaluating the short-term exposure studies to help inform consideration of
15    the protection afforded by an annual PM2.5 standard, we observe, based on quantitative risk
16    assessments conducted in past reviews, that much of the risk related  to daily exposures, when
17    aggregated on  an annual basis, results from the large number of days during which the 24-hour
18    average concentrations are in the low- to mid-range of the entire distribution. Thus, to reduce the
19    aggregate short-term exposure-related risk, it is necessary to shift the bulk of the air quality
20    distribution to  lower levels, not just to limit the concentrations on days when the PM2 5
21    concentrations are  relatively high.
22           While shifting the distribution can be accomplished through control strategies aimed at
23    meeting either an annual or 24-hour standard, we have seen in quantitative risk assessments
24    conducted in this and past reviews that more consistent aggregated risk reductions across study
25    areas are likely to be accomplished through strategies aimed at meeting an annual standard.
26    Thus, in considering the adequacy of the current suite of standards in protecting against effects
27    associated with short-term exposures, we consider whether associations in short-term exposure
28    studies have been reported in areas in which the aggregate long-term study mean concentrations
29    are at or below the level of the current annual standard, as well as by considering whether the
30    98th percentile PM2 5 concentrations in such studies are at or below the level of the current 24-
31    hour standard.  We conclude that such short-term exposure studies can reasonably be viewed as
32    calling into question the adequacy of the current annual and/or 24-hour standards.
      4 While this view is consistent with that presented in the 2005 Staff Paper, in the review completed in 2006, as noted
      above, the Administrator chose to focus more narrowly and considered the short-term exposure evidence as a basis
      for the 24-hour standard alone.
             Draft Do Not Quote or Cite           2-11                         March 2010

-------
 1           In considering what alternative standards would be protective against effects observed in
 2    such short-term exposure studies, we first note the absence of any discernable threshold within
 3    the distribution of 24-hour average concentrations in the short-term exposure studies. While
 4    recognizing that health effects may occur over the full range of 24-hour concentrations observed
 5    in the studies, as with the long-term exposure studies, we believe that it is reasonable to conclude
 6    that the evidence of association is strongest down to somewhat below the mean 24-hour average
 7    concentration,  such as down to one standard deviation below the mean  or to the lower end of the
 8    interquartile range, where the data in the study are most concentrated.  We also believe it is
 9    appropriate to consider whether, and if so, where in the distribution of 24-hour average
10    concentrations the confidence interval  becomes notably wider, suggestive of a concentration
11    below which the association becomes appreciably more uncertain and the possibility that an
12    effects threshold may exist becomes more likely.  Based on these considerations, we identify a
13    range of alternative annual standard levels that we conclude are appropriate to consider for
14    protecting against PM2.5-related short-term exposure effects based on the evidence available in
15    this review.  We also recognize that it is unlikely that an annual standard could effectively
16    protect against short-term exposure-related effects in all areas across the country, especially in
17    areas with relatively high peak-to-mean ratios in PM2.5 concentrations or in areas that are
18    strongly affected by localized "hotspots" or by seasonal exposures that  may not be well
19    controlled by an annual standard. Thus, we also identify a range of alternative 24-hour standard
20    levels that we judge to be appropriate to consider for protecting against PM2.5-related short-term
21    exposure effects, in conjunction with a range of alternative annual standards, based on the
22    evidence available in this review.
23           Based on the evidence-based considerations outlined above, we then develop integrated
24    preliminary conclusions with regard to alternative suites of standards, including both annual and
25    24-hour standards, that we believe are  appropriate for consideration in this review based on the
26    currently available evidence. In so doing, we discuss the roles that each standard might be
27    expected to play in the protection afforded by alternative suites of standards.
28           Beyond these evidence-based considerations, we also consider the quantitative risk
29    estimates and the key observations presented in the second draft quantitative risk assessment
30    (RA). This assessment included an evaluation of 15 urban case study areas and estimated risk
31    associated with a number of health endpoints associated with long-term and short-term PM2.5
32    exposures (US EPA,  2010a). As part of our risk-based considerations,  we have considered
33    estimates of the magnitude of PM2.5-related risks associated with recent air quality levels and air
34    quality simulated to just meet the current and alternative suites of standards using alternative
35    simulation approaches. We have also characterized the risk reductions, relative to the risks
36    remaining upon just meeting the current standards, associated with just meeting alternative suites
             Draft Do Not Quote or Cite           2-12                       March 2010

-------
 1    of standards. In so doing, we recognize the uncertainties inherent in such risk estimates, and
 2    have taken such uncertainties into account by considering the sensitivity of the "core" risk
 3    estimates to alternative assumptions and methods likely to have substantial impact on the
 4    estimates. In addition, we have conducted additional analyses to characterize the
 5    representativeness of the urban study areas within a broader national context. We have
 6    considered this risk-based information to help inform our preliminary conclusions on the
 7    adequacy of the current suite of standards, potential alternative suites of standards that are
 8    appropriate for consideration in this review, and on the roles that the annual and 24-hour
 9    standards may play in affording protection against effects related to both long- and short-term
10    PM2.5 exposures.
11           Our preliminary conclusions reflect our understanding of both evidence-based and risk-
12    based considerations to inform two overarching questions related to: (1) the adequacy of the
13    current suite of PM2.5 standards and (2) what potential alternative standards, if any, should be
14    considered in this review to provide appropriate protection from the effects associated with both
15    long- and short-term exposures to fine particles.  In addressing these broad questions, we have
16    organized the discussions below around a series of more specific questions reflecting different
17    aspects of each overarching question.  When evaluating the health protection afforded by the
18    current or any alternative suites of standards considered, we have taken into account the four
19    basic elements of the NAAQS (e.g., indicator, averaging time, form, and level).
20           We believe that the approach outlined above, when presented in the final PA, will
21    provide a comprehensive basis to help inform the judgments required of the Administrator in
22    reaching decisions about the current and potential alternative primary PM2 5 standards and in
23    responding to the remand of the 2006 annual PM2.5 standard.

24    2.2     ADEQUACY OF CURRENT STANDARDS
25           In considering the adequacy of the current suite of PM2 5 standards, staff addresses the
26    following overarching question:
27    Does the currently available scientific evidence and risk-based information, as reflected in
28     the ISA and RA, support or call into question the adequacy of the protection afforded by
29   	the current suite of fine particle standards?	
30           To inform the answer to this broad question, we have posed  a series of more  specific
31    questions to aid in considering the currently available scientific evidence and the results of recent
32    quantitative risk analyses in a policy-relevant context, as discussed below. In considering the
33    scientific and technical information, we reflect upon both the information available in the last
34    review  and information that is newly available since the last review as assessed and presented in
3 5    the ISA and second draft RA (US EPA, 2009a; US EPA, 201 Oa).

             Draft Do Not Quote or Cite           2-13                         March 2010

-------
 1    2.2.1  Evidence-based Considerations
 2          In reviewing the adequacy of the current suite of PM2.5 standards, we have taken into
 3    account evidence-based considerations primarily by assessing the currently available evidence as
 4    presented in the ISA.  Our review of the adequacy of the current standards begins by considering
 5    causal inference, impacts on susceptible populations, and whether health effects have been
 6    observed in areas where the air quality concentrations extend to levels lower than previously
 7    reported.
 8    •   To what  extent does the newly available scientific evidence and related uncertainties
 9       strengthen or call into question evidence of associations between ambient fine particle
10       exposures and health effects?
11          In considering the strength of the associations between short- and long-term exposures to
12    PM2.5 and health effects, we first recognize that,  in the last review, EPA concluded that there was
13    "strong epidemiological evidence" for PM2.5 linking short-term  exposures with cardiovascular
14    and respiratory mortality and morbidity, and long-term exposures with cardiovascular and lung
15    cancer mortality and respiratory morbidity (US EPA, 2004, p. 9-46; US EPA, 2005, p. 5-4).
16    Overall, the epidemiological evidence supported "likely causal associations" between PM2 5 and
17    both mortality and morbidity from cardiovascular and respiratory diseases, based on "an
18    assessment of strength, robustness, and consistency in results" (US  EPA, 2004, p. 9-48).
19          In looking across the extensive new  scientific evidence available in this review, our
20    overall understanding of health effects associated with fine particle exposures has been greatly
21    expanded. The currently available evidence is stronger in comparison to evidence  available in
22    the last review because of its breadth and the substantiation of previously observed health effects
23    (US EPA, 2009a, section 2.3.1).  A number of large multi-city epidemiological studies have been
24    conducted throughout the U.S. including extended analyses of studies that were important to
25    inform decisionmaking in the last review. These studies have reported consistent increases in
26    morbidity and/or mortality related to ambient PM2.5 concentrations, with the strongest evidence
27    reported for cardiovascular-related effects.  In addition, the findings of new toxicological and
28    controlled human exposure studies provide  stronger support for a number of potential biologic
29    mechanisms  or pathways for PM-related cardiovascular and respiratory effects (US EPA, 2009a,
30    chapter 5; Figures 5-4 and 5-5).  In summary, the ISA concludes, "[t]he new evidence ...  greatly
31    expands upon the evidence available in the 2004 PM AQCD particularly in providing greater
32    understanding of the underlying mechanisms for PM2 5 induced  cardiovascular and respiratory
33    effects for both short- and long-term exposures" (US EPA, 2009a, p. 2-17).
34          As an initial matter, we note that since the conclusion of the last PM NAAQS review, the
35    Agency has developed a more formal framework for reaching causal inferences from the body of
36    scientific evidence drawing upon the evaluation  and synthesis of evidence from across
            Draft Do Not Quote or Cite           2-14                        March 2010

-------
 1    epidemiological, controlled human exposure, and toxicological studies.  This framework uses a
 2    five-level hierarchy that classifies the overall weight of evidence and causality using the
 3    following categorizations: causal relationship, likely to be a causal relationship, suggestive of a
 4    causal relationship, inadequate to infer a causal relationship, and not likely to be a causal
 5    relationship (US EPA, 2009a, Table 1-3).
 6           In reaching causal inferences, EPA has considered uncertainties that bear on our
 7    understanding of the body of currently  available scientific evidence (US EPA, 2009a, section
 8    1.5). For example, in epidemiological  studies the potential for confounding bias remains an
 9    important source of uncertainty in evaluating the health effects associated with one pollutant that
10    is part of a larger, complex mixture of pollutants. Epidemiological studies attempt to disentangle
11    the effects of the air pollution mixture and identify the health effects associated with a specific
12    pollutant, such as PM2.5, using multivariate regression models to control for the potential
13    confounding effects by other pollutants (e.g., gaseous co-pollutants5) for which measurements
14    are available.  However, there are several statistical issues influencing results generated using co-
15    pollutant models that lead to uncertainty in the quantitative interpretation of these results. In this
16    review, the uncertainties associated with the evaluation of the body of scientific evidence for
17    PM2.5-associated health effects remain  largely the same as in previous reviews (e.g., co-pollutant
18    confounding, exposure misclassification) and have been considered in reaching causality
19    determinations as discussed in the ISA.
20           Looking broadly to integrate epidemiological evidence with that from controlled human
21    exposure studies and toxicologic studies and using  this causal framework, the ISA concludes that
22    the collective evidence is largely consistent with past studies and substantially  strengthens what
23    was known in the last review to reach the conclusion that a causal relationship exists between
24    both short- and long-term exposures to PM2.5 and mortality and cardiovascular effects.
25    Furthermore, the ISA concludes that the collective  evidence continues to support likely causal
26    associations between short- and long-term PM2.5 exposures and respiratory effects.  Additional
27    evidence is suggestive of a causal relationship between long-term PM2.5 exposures and other
28    health effects, including developmental and reproductive effects (e.g., low birth weight) and
29    carcinogenic, mutagenic, and genotoxic (e.g., lung  cancer mortality) effects. Table 2-1
30    summarizes the causal determinations for health outcomes associated with short- and long-term
31    exposures to PM2.5 (US EPA, 2009a, sections 2.3.1  and 2.6).
32
      5 A gaseous co-pollutant (e.g., O3, CO, SO2, NO2) meets the criteria for potential confounding in PM2 5-related
      effects if: (1) it is a potential risk factor for the health effect under study; (2) it is correlated with PM2 5 and (3) it
      does not act as an intermediate step in the pathway between PM2 5 exposure and the health effect under study (US
      EPA, 2004, p. 8-10).
             Draft Do Not Quote or Cite           2-15                         March 2010

-------
                        Table 2-1. Summary of Causality Determinations for PM2.s
Exposure
Duration
Long-term
Short-term
Outcome
Mortality
Cardiovascular Effects
Respiratory Effects
Reproductive and Developmental Effects
Cancer, Mutagenicity, Genotoxicity Effects
Mortality
Cardiovascular Effects
Respiratory Effects
Central Nervous System Effects
Causality
Determination
Causal
Causal
Likely to be Causal
Suggestive
Suggestive
Causal
Causal
Likely to be Causal
Inadequate
 2    Source:  adapted from US EPA, 2009, Table 2-6.
 3    Health Effects Associated with Long-term PM2 s Exposure
 4          With regard to mortality associated with long-term PM2.5 exposures, the ISA concludes
 5    that newly available evidence significantly strengthens the evidence linking long-term exposure
 6    to PM2.5 and mortality, while providing indications that the magnitude of the PM2.5-mortality
 7    association may be larger than previously estimated (US EPA, 2009a, sections 7.2.10, 7.2.11,
 8    7.6.1; Figures 7-6 and 7-7). A number of large U.S. cohort studies have been published since the
 9    last review. This evidence includes new analyses and insights from extended analyses of the
10    American Cancer Society (ACS) and Harvard Six Cities studies (US EPA, 2009a, pp 7-84 to 7-
11    85; Figure 7-6; Krewski et al., 2009; Pope et al., 2004; Jerrett et al., 2005; Laden et al., 2006).  In
12    addition, new long-term PM2 5 exposure studies evaluating mortality impacts in additional
13    cohorts are now available. The Women's Health Initiative has investigated the effects of PM2.5
14    on cardiovascular-related mortality in post-menopausal women with no previous history of
15    cardiac disease (US EPA, 2009a, p. 7-87; Miller et al., 2007). Mortality impacts in older adults
16    using Medicare data have been reported by a number of investigators (US EPA, 2009a, pp. 7-87
17    to 7-89; Eftim et al., 2008; Zeger et al., 2007, 2008). We note that other long-term PM2.5
18    exposure studies provide additional evidence of mortality associated with long-term PM2.5
19    exposures as discussed more fully in section 7.6 of the ISA.  Collectively, these long-term PM2 5
20    exposure studies, along with the evidence available in the last review, provide us with consistent
21    and stronger evidence of associations between long-term exposure to PM2.5 and mortality (U.S.
22    EPA, 2009a, sections 2.3.1  and 7.6).
            Draft Do Not Quote or Cite
2-16
March 2010

-------
 1          When integrating the mortality evidence, EPA considered impacts on all-cause or non-
 2    accidental mortality as well as cause-specific mortality (e.g., cardiovascular-related mortality,
 3    respiratory-related mortality, lung cancer mortality) and reached the conclusion that there is a
 4    causal association between mortality and long-term PM2.5 exposure (US EPA, 2009a, pp. 2-12
 5    and 7-96). The strongest evidence indicates an association between long-term PM2.5 exposures
 6    and mortality due to cardiovascular disease, with additional evidence supporting an association
 7    between PM2.5 and mortality related to lung cancer (US EPA, 2009a, pp. 2-12, 7-81, and 7-95 to
 8    7-96; Figure 7-7). We note that fewer long-term PM2.5 exposure studies have evaluated the
 9    respiratory component of cardiopulmonary mortality, and the evidence to support an association
10    with long-term exposure to PM2.5 and respiratory mortality is, therefore, much more limited (US
11    EPA 2009a, p. 2-12; Figure 7-7).  The strength of associations for cause-specific mortality is
12    coherent with cardiovascular-related morbidity endpoints as discussed below. The most recent
13    evidence for cardiovascular-related mortality in post-menopausal women indicates that the
14    evidence of association with long-term exposure to PM2.5 in this population is "particularly
15    strong" (US EPA, 2009a, p. 7-96).
16          In addition, the strength of the causal association between long-term PM2.5 exposure and
17    mortality builds upon new studies providing evidence of improvement in community health
18    following reduction in ambient fine particle concentrations.  Pope et al. (2009) have documented
19    the population health benefits of reducing ambient air pollution by correlating past reductions in
20    ambient PM2.5 concentrations with increased life expectancy. These investigators report that
21    reductions in fine particle ambient concentrations that occurred during the 1980s and 1990s
22    account for as much as 15 percent of the overall improvement in life expectancy in  51 U.S.
23    metropolitan areas also analyzed in the ACS study. The decrease in fine particle exposure was
24    reported to be associated with an estimated increase in mean life expectancy of approximately 5
25    to 9 months (US EPA, 2009a, p. 7-95; Pope et al., 2009). An extended analysis of the Harvard
26    Six Cities study found that as cities cleaned up their air, locations with the largest reductions in
27    PM2.5 saw the largest improvements in reduced mortality rates, while those with the smallest
28    decreases in PM2.5 concentrations saw the smallest improvements in reduced mortality rates
29    (Laden et al., 2006). Reduced mortality risk observed in this extended follow-up study was
30    related to deaths due to cardiovascular and respiratory-related disease, but not from lung cancer
31    (US EPA, 2009a, p. 7-84). An additional extended follow-up to the Harvard Six Cities study
32    investigated the delay between changes in exposure and changes in mortality.  Schwartz et al.
33    (2008) reported that the effects of changes in PM2.5 exposures were seen within the 2 years prior
34    to death (US EPA, 2009a, p. 7-92; Figure 7-9).
35          With regard to cardiovascular effects associated with fine particle exposures, the ISA
36    includes consideration of both cardiovascular-related mortality as well as morbidity effects and
            Draft Do Not Quote or Cite           2-17                        March 2010

-------
 1    concludes that a causal association exists between long-term PM2.5 exposure and cardiovascular
 2    effects (US EPA, 2009a, pp. 2-12 and 7-19). Recent studies have provided new evidence linking
 3    long-term exposure to PM2 5 with cardiovascular outcomes that has "expanded upon the
 4    continuum of effects ranging from the more subtle subclinical measures to cardiopulmonary
 5    mortality" (US EPA, 2009a, p. 2-17).  Several new epidemiologic studies have examined the
 6    association between cardiovascular effects and long-term PM2.5 exposures in multi-city studies
 7    conducted in the U.S. and Europe.  The ISA concludes that the strongest evidence of
 8    cardiovascular effects related to long-term exposure to PM2 5 has been reported in recent studies
 9    investigating cardiovascular-related mortality. This includes evidence from a number of large,
10    multi-city U.S. long-term cohort studies including extended follow-up analyses of the ACS and
11    Harvard Six Cities studies as well as the WHI as outlined above (US EPA, 2009a, sections 7.2.10
12    and 7.6.1; Krewski  et al., 2009; Pope et al., 2004; Laden et al., 2006; Miller et al., 2007).  Pope
13    et al. (2004) report a positive association between mortality and long-term PM2 5 exposure for a
14    number of specific cardiovascular diseases, including ischemic heart disease (IHD),
15    dysrhythmia, heart failure, and cardiac arrest (US EPA, 2009a, p. 7-84; Figure 7-7). Krewski et
16    al. (2009) provides  further evidence and precision for IHD-related mortality associated with
17    long-term PM2.5 exposure (US EPA, 2009a, p. 7-84, Figure 7-7).
18          Epidemiologic studies examining cardiovascular morbidity associated with long-term
19    PM2 5 exposures were not available in the previous PM reviews. In the current review, studies
20    are now available that evaluated  a number of endpoints ranging from subtle indicators of
21    cardiovascular health to serious clinical events associated with coronary heart disease (CHD) and
22    cerebrovascular disease (CVD) including myocardial infarction (MI), coronary artery
23    revascularization (e.g., bypass graft, angioplasty,  stent, atherectomy), congestive heart failure
24    (CHF), and stroke.  The most significant new evidence comes from the WHI study which
25    provides evidence of nonfatal cardiovascular events including both coronary and cerebrovascular
26    events in a cohort of post-menopausal women with no previous history of cardiac disease (Miller
27    et al., 2007; US EPA, 2009a, sections 7.2.9 and 7.6.1).
28          As noted in  the ISA, there may be multiple mechanisms related to the observed
29    associations between PM2 5 and cardiovascular effects, and these processes may be interlinked
30    (US EPA, 2009a, sections 5.2,  5.3, 5.5, 5.6, 5.7).  For example, myocardial ischemia and MI may
31    occur as a result the proposed effects of PM on atherosclerosis, plaque instability, plaque rupture,
32    thrombosis, and/or altered vasoreactivity of coronary vessels. Myocardial ischemia and MI may
33    alter the conduction and depolarization properties of the heart and lead to arrhythmic events. In
34    addition, thrombosis may lead to stroke and/or thromboembolic disease.  The ISA notes that "it
35    is not clear at this time whether PM initiates cardiovascular disease or whether it perturbs
36    existing disease" (US EPA, 2009a, p. 5-18).
            Draft Do Not Quote or Cite           2-18                        March 2010

-------
 1           Toxicological studies provide evidence that the cardiovascular morbidity effects observed
 2    in epidemiological studies are biologically plausible and coherent with studies of cardiovascular-
 3    related mortality observed in long-term exposures as well as with studies of cardiovascular-
 4    related morbidity and mortality associated with short-term exposures to PM2.5 as described below
 5    (US EPA, 2009a, p 7-19). For example, it has been hypothesized that that exposure to PM2.5 can
 6    lead to myocardial ischemia through an effect on the autonomic nervous system or by altering
 7    vasomotor function. Furthermore, PM-induced systemic inflammation and oxidative stress may
 8    contribute to altered vasomotor function which has been demonstrated as altered microvascular
 9    reactivity, altered vessel tone, and reduced blood flow during ischemia.6  Toxicological studies
10    demonstrating increased right ventricular pressure and diminished cardiac contractility also
11    provide biological plausibility for the associations observed between PM2 5 and CFIF in
12    epidemiological studies (US EPA, 2009a, p. 2-15). Additional evidence of cardiovascular-
13    related morbidity and mortality associated with long-term PM2.5 exposures is  discussed more
14    fully in section 7.2 of the ISA.  Additional evidence discussing potential mechanisms underlying
15    cardiovascular effects  are discussed more fully in sections 5.2 through 5.6 of the ISA.
16           The ISA notes extended analyses of studies available in the last review as well as new
17    epidemiologic studies  conducted in the U.S. and abroad provide stronger evidence of respiratory-
18    related morbidity associated with long-term PM2.5 exposure. As mentioned above, more limited
19    data are available for respiratory-related mortality effects.  Considering morbidity and mortality
20    effect collectively, the ISA has concluded that there is continued support for a likely causal
21    association between long-term PM2.5 exposures and respiratory effects. The strongest evidence
22    for respiratory-related effects available in this review is from studies that have evaluated
23    decrements in lung function growth,  increased respiratory symptoms, and asthma development
24    (U.S. EPA, 2009a, sections 2.3.1.2, 7.3.1.1, and 7.3.2.1).7  Specifically, extended analyses of the
25    Southern California Children's Health  Study (CHS) provide evidence that clinically important
26    deficits in lung function8 associated with long-term exposure to PM2.5 persisted into early
      6The vasculature of all tissues is lined with endothelial cells that will naturally encounter any systemically absorbed
      toxin.  The endothelium (1) maintains barrier integrity to ensure fluid compartmentalization, (2) communicates
      dilatory and constrictive stimuli to vascular smooth muscle cells, and (3) recruits inflammatory cells to injured
      regions. Smooth muscle cells lie within the layer of endothelium and are crucial to the regulation of blood flow and
      pressure. In states of injury and disease, both cell types can exhibit dysfunction and even pathological responses.
      Endothelial dysfunction is a factor in many diseases and may contribute to the origin and/or exacerbation of
      perfusion-limited diseases, such as MI or IHD, as well as hypertension. Endothelial dysfunction is also a
      characteristic feature of early and advanced atherosclerosis (US EPA, 2009a, p. 6-24).
      7 Supporting evidence comes from studies "that observed associations between long-term exposure to PM10 and an
      increase in respiratory symptoms and reductions in lung function grown in areas where PM10 is dominated by PM2 5"
      (US EPA, 2009a, p. 2-12).
      o
       Clinical significance was defined as a FEVi below 80% of the predicted value, a criterion commonly used in
      clinical settings to identify persons at increased risk for adverse respiratory conditions (US EPA, 2009a, p. 7-29-7-
      30).
             Draft Do Not Quote or Cite            2-19                          March 2010

-------
 1    adulthood (U.S., EPA, 2009a, p. 7-27; Gauderman et al., 2004).  Additional analyses of the CHS
 2    cohort report an association between long-term PM2.5 exposure and bronchitic symptoms (US
 3    EPA, 2009a, p. 7-24; McConnell et al., 2003) and a strong modifying effect on the association
 4    between lung function with asthma incidence (US EPA, 2009, 7-25; Islam et al., 2007). The
 5    strength and robustness of the outcomes observed in these more recent reports from the Southern
 6    California CHS, including evaluation of a broader range of endpoints than previous CHS studies
 7    with shorter follow-up periods, were larger in magnitude and more precise. Supporting these
 8    results are new longitudinal cohort studies conducted by other researchers in varying locations
 9    using different methods (U.S. EPA, 2009a, section 7.3.9.1).  New evidence from a U.S. cohort of
10    cystic fibrosis (CF) patients provides evidence of association between long-term PM2.5 exposures
11    and exacerbations of respiratory symptoms resulting in hospital admissions or use of home
12    intravenous antibiotics (US  EPA, 2009a,  p. 7-25; Goss et  al., 2004).
13          Toxicological studies provide coherence and biological plausibility for the respiratory
14    effects observed in epidemiological studies (US EPA,  2009a, p. 7-42). For example, pre- and
15    postnatal exposure to ambient levels of urban particles has been found to affect lung
16    development in an animal model (US EPA, 2009a, section 7.3.2.2; p. 7-43).  This finding is
17    important because impaired lung development is one mechanism by which PM exposure may
18    decrease lung function growth in children (US EPA, 2009a, p. 2-12; section 7.3). Subchronic
19    and chronic toxicological studies of concentrated ambient particles (CAPs) as well as evaluations
20    of specific sources of fine particles including, diesel exhaust (DE), roadway air, and woodsmoke
21    provide some evidence for altered pulmonary function, mild inflammation, oxidative responses,
22    immune suppression, and histopathologic changes. In addition, exacerbated allergic responses
23    have been observed in animals exposed to DE and wood smoke (US EPA, 2009a, p. 2-12,
24    section 7.3).
25          With respect to respiratory-related mortality associated with long-term PM2.5 exposure,
26    evidence is "limited and inconclusive" (US EPA, 2009a, p. 7-41). The extended follow-up of the
27    Harvard Six Cities study reports a positive but non-statistically significant association between
28    long-term PM2.5 exposure and respiratory-related mortality (Laden et al., 2006). Pope et al.
29    (2004) found no association with long-term PM2.5 exposure  and respiratory-related mortality (US
30    EPA, 2009a, p. 7-84). There is emerging but limited evidence for an association between long-
31    term PM2.5 exposure and respiratory mortality in post-neonatal infants where long-term exposure
32    was considered as approximately one month to one year (US EPA, 2009a, pp. 7-54 to 7-55).
33    Emerging evidence of short- and long-term exposure to PM2.5 and respiratory morbidity effects
34    and infant mortality are coherent with the weak respiratory mortality effects observed.
35          Beyond effects considered to have causal or likely causal associations with long-term
36    PM2.s exposure as discussed above, the ISA also notes health outcomes classified as having
            Draft Do Not Quote or Cite           2-20                        March 2010

-------
 1    evidence suggestive of a causal association with long-term PM2.5 exposure.  This includes two
 2    broad categories of health outcomes: (1) reproductive and developmental effects and (2) cancer,
 3    mutagenicity, and genotoxicity effects (US EPA, 2009a, Table 2-6). With respect to
 4    reproductive and developmental effects, the ISA notes, "[e]vidence is accumulating for PM2.5
 5    effects on low birth weight and infant mortality, especially due to respiratory causes during the
 6    post-neonatal period" (US EPA, 2009a, section 2.3.1.2). New evidence available in this review
 7    reports a significant association between exposure to PM2.5 during pregnancy and lower birth
 8    weight, pre-term birth, and intrauterine growth restriction, respectively, and post-natal exposure
 9    to PM2.5 associated with an increased risk of infant mortality (US EPA, 2009a, section 7.4). The
10    ISA further notes that "[i]nfants and fetal development processes may be particularly vulnerable
11    to PM exposure, and although the physical mechanisms are not fully understood, several
12    hypotheses have been proposed involving direct effects on fetal health, altered placenta function,
13    or indirect effects on the mother's health" (US EPA, 2009a, section 7.4.1). However,
14    toxicological studies provide some evidence which is coherent with an association between long-
15    term PM2.5 exposure and adverse reproductive and developmental outcomes, but provide  "little
16    mechanistic information or biological plausibility for an association between long-term PM
17    exposure and adverse birth outcomes (e.g., low birth weight or infant mortality)" (US EPA,
18    2009a, p. 2-13).
19          With respect to cancer, mutagenic and genotoxic effects, "[mjultiple epidemiologic
20    studies have shown a consistent positive association between PM2.5 and lung cancer mortality,
21    but studies have generally not reported associations between PM2 5 and lung cancer incidence"
22    (US EPA, 2009a, sections 2.3.1.2 and 7.5; Table 7-7; Figures 7-6 and 7-7).  The extended
23    follow-up to the ACS study reported an association between long-term PM2.5 exposure and lung
24    cancer mortality (US EPA, 2009a, p. 7-71; Krewski et al., 2009). The extended follow-up of the
25    Harvard Six Cities study also evaluated lung cancer mortality and reported a positive association
26    when considering the entire 25-year follow-up period.  However, estimated decreases in average
27    PM2.s concentrations between the first and second follow-up periods were not associated with
28    reduced  lung cancer mortality (US EPA, 2009a, p. 7-71; Laden et al, 2006).  Epidemiological
29    evidence is not currently available to evaluate cancer in organs or systems other than the lung
30    related to long-term exposure to PM2 5 (US EPA, 2009a, p. 7-81).  There is some evidence,
31    primarily from in vitro studies, providing biological plausibility for the PM-lung cancer
32    relationships observed in epidemiological studies (US EPA, 2009a, p. 7-80). Toxicological
33    studies providing evidence of carcinogenicity, mutagenicity, and genotoxicity reported mixed
34    results as discussed in section 7.5 of the ISA.
35    Health Effects Associated with Short-term PMi.s Exposure

            Draft Do Not Quote or Cite           2-21                        March 2010

-------
 1           In considering effects associated with short-term PM2.5 exposure, the body of currently
 2    available scientific evidence has been expanded greatly by the publication of a number of new
 3    multi-city time-series studies that have used uniform methodologies to investigate the effects of
 4    short-term fine particle exposures on public health.  This body of evidence provides a more
 5    expansive data base and considers multiple locations representing varying regions and seasons
 6    that provide evidence on the influence of climate and air pollution mixes on PM2.5-associated
 7    health effects.  These studies provide more precise estimates of the magnitude of effects
 8    associated with short-term PM2.5 exposure than most smaller-scale single-city studies that were
 9    more commonly available in the last review (U.S. EPA 2009a, chapter 6).
10           With regard to mortality effects, looking broadly across all of the available scientific
11    evidence, the ISA concludes that a causal association exists between short-term PM2 5 exposure
12    and premature death. Extended and expanded analyses of multi-city studies available in the last
13    review as well as a large number of new U.S. multi-city and single-city short-term PM2.5
14    exposure studies have found generally consistent positive associations between short-term PM2.5
15    exposures and cardiovascular- and respiratory-related mortality as well  as all-cause (non-
16    accidental)  mortality (US EPA, 2009a, sections 2.3.1.1, 6.2.11 and 6.5.2.2; Figures 6-26, 6-27,
17    and 6-28).   These newly published U.S. multi-city short-term PM2.5 exposure studies provide a
18    much larger body of evidence since the last review. In an analysis of National Morbidity,
19    Mortality, and Air Pollution Study (NMMAPS) data, Dominici et al. (2007) reported
20    associations between fine particle  exposures and all-cause and cardio-respiratory mortality (US
21    EPA, 2009a, p. 6-175, Figure 6-26). A Canadian multi-city study available in the last review
22    was expanded from 8 to 12  cities (Burnett and Goldberg, 2003; Burnett et al., 2004). The results
23    of this expanded analysis were consistent with the earlier findings of a positive and statistically
24    significant association between short-term PM2 5 exposure and mortality (US EPA, 2009a, p 6-
25    182, Figure 2-1). However,  the influence of NO2 and limited PM2.5 data for several years during
26    the study period somewhat diminish these findings and underscore the need for additional data
27    on the co-pollutants in relation to the PM2.5-mortality association. Zanobetti and Schwartz
28    (2009) reported positive (99%) and frequently statistically significant (55%) associations across
29    most of the 112 cities between short-term PM2 5 exposure and total, cardiovascular-related, and
30    respiratory-related mortality (US EPA, 2009a, pp 6-176 to 6-179; Figures 6-23 and 6-24).
31    Collectively, these studies provide generally consistent and much stronger evidence than the
32    evidence available in the last review.
33           With respect to cardiovascular effects, the ISA considers both cardiovascular-related
34    mortality as well as cardiovascular morbidity effects in reaching the conclusion that there is a
35    causal association between short-term PM2 5 exposure and cardiovascular effects. New multi-
36    city as well as single-city short-term PM2.5 exposure studies conducted since the prior review
             Draft Do Not Quote or Cite           2-22                        March 2010

-------
Figure 2-1. Summary of Effect Estimates (per 10 iig/m ) and Air Quality Distributions for Long-term Exposure Studies
Study
Outcome
Mean
.(SD)
Zegeretal. (2008) All-Cause Mortality, Central U.S. 10.7*
Kim et al. (2004) Bronchitis (Children) 1 2.0
Zegeretal. (2008) All-Cause Mortality, Western U.S. 13.1*
Miller et al. (2007) CVD Morbidity or Mortality 1 3.5 (3.3)
Eftimetal. (2008) All-Cause Mortality, ACS Sites 13.6(2.8)
Goss et al. (2004) All-Cause Mortality 1 3.7 (4.2)
McConnell et al. (2003) Bronchitis (Children) 1 3.8 (7.7)
Zeger et al. (2008) All-Cause Mortality, Eastern U.S. 1 4.0*
Krewski et al. (2009) All-Cause Mortality 1 4.0 (3.0)
Eftimetal. (2008) All-Cause Mortality, Harv 6-Cities 14.1(3.1)
Lipfert et al. (2006) All-Cause Mortality 1 4.3 (3.0)
Dockery et al. (1 996) Bronchitis (Children) 1 4.5 (4.2)
Woodruff etal. (2008) Infant Mortality (Respiratory) 14.9****
Laden et al. (2006) All-Cause Mortality 1 6.4 (5.6)*
Woodruff etal. (2006) Infant Mortality (Respiratory) 19.2
Enstrom (2005) All-Cause Mortality 23.4
Chen etal. (2005) CHD Mortality, Females 29.0
CHD Mortality, Males 29.0
* Zeger et al. (2008) reported median with IQR; overall study reported median of 1 3
Mean
-1SD
Lower IQR
9.8
10.4
10.2 11.6
10.8
9.5 11.8
6.1
12.3
11.0 11.8
11.0
11.3
10.3
12.0
10.8
2 pg/m3( lower IQR -11.1)
Range Effect Estimate (95% Cl)
9.8-12.2*
11-15***
10.4-18.5* -J
34 283
6.0-25.1
11 8 159
6-29
12.3-15.3*
5.8-22
9.6-19.1
5.0 -?
j
10-22


-—
L
-

u-

l_ —
L_


1 1 ' 1 1 I 1 1 1
0.7 0.9 1-1 1-3 1-5 1.7 1.9 2.1
"Estimated from data provided by study author
""range of averages at 10 schools
**** median for all cause mortality; median for survivors; = 14.8, lower end of IQR (11.7)
Modified from Figure 2-2, US EPA, 2009a.
                                                                                                              Relative Risk / Odds Ratio

-------
 1    support a largely positive and frequently statistically significant relationship between short-term
 2    exposure to PM2.5 and cardiovascular-related disease, substantiating prior findings of a positive
 3    relation between exposure to fine particles and cardiovascular morbidity. For example, short-
 4    term exposure to PM2.5 in association with cardiovascular and respiratory effects was evaluated
 5    among a multi-city cohort of older adults participating in the Medicare Air Pollution Study
 6    (MCAPS) (US EPA, 2009a, pp. 6-57 to 58; Dominici et al, 2006a; Bell et al, 2008). Overall,
 7    short-term PM2.5 exposure studies available in the current review provide consistent evidence of
 8    a positive association between short-term PM2.5 exposures and hospital admissions (HA) or
 9    emergency department (ED) visits as well as premature mortality related to cardiovascular
10    outcomes. The strongest evidence has been observed for cardiovascular morbidity effects
11    predominately associated with HA and ED visits reported for ischemic heart disease (IHD) and
12    congestive heart failure (CHF) and cardiovascular-related mortality (US EPA, 2009a, Figure 2-1,
13    p. 6-79, sections 6.2.10.3, 6.2.10.5, and 6.2.11; Bell et al., 2008; Dominici et al., 2006a; Tolbert
14    et al., 2007; Zanobetti and Schwartz, 2009).  Furthermore, these findings are supported by a
15    recent study of a multi-city cohort of women participating in the Women's Health Initiative that
16    reported a non-significantly association between short-term exposure to PM2.5 and myocardial
17    ischemia (Zhang et al., 2009).
18          In focusing on respiratory effects, the ISA integrates evidence of respiratory-related
19    mortality with respiratory morbidity effects to reach the conclusion that there is a likely causal
20    association between short-PM2.5 exposure and respiratory  effects. The strongest evidence from
21    short-term PM2.5 exposure studies has been observed for respiratory-related ED visits and HAs
22    for chronic obstructive pulmonary disease (COPD) and respiratory infections (U.S. EPA, 2009a,
23    sections 2.3.1.1 and 6.3.8.3; Figures 2-1 and 6-13; Dominici et al., 2006a;).  Evidence for PM2.5-
24    related respiratory effects has also been observed in panel studies, which indicate associations
25    with respiratory symptoms, pulmonary function, and pulmonary inflammation among asthmatic
26    children.  Although not consistently observed, some controlled human exposure studies have
27    reported small decrements in various measures of pulmonary function following controlled
28    exposures to PM2.5 (US EPA, 20009a, p. 2-10).  Furthermore, the comparatively larger body of
29    toxicological evidence since the last review is coherent with short-term exposures to PM2.5 and
30    respiratory effects (US EPA 2009a, section 6.3.10.1).
31
32    Summary
33          In considering the extent to which newly available scientific evidence strengthens or calls
34    into question evidence of associations identified in the last review between ambient fine particle
35    exposures and health  effects, we recognize that much progress has been made in assessing some
36    key uncertainties related to our understanding of health effects associated with short- and long-
            Draft Do Not Quote or Cite          2-24                        March 2010

-------
 1    term exposure to PM2.5. As briefly discussed above as well as in the more complete discussion of
 2    the evidence as assessed in the ISA, we note that the newly available information combined with
 3    information available in the last review provides substantially stronger confidence in a causal
 4    association between short- and long-term exposures to PM2.5 and mortality and cardiovascular
 5    effects.  In addition, the newly available evidence reinforces and expands the evidence
 6    supporting the likely causal nature of the associations between short- and long-term exposure to
 7    PM2.s and respiratory effects.  Causal inferences, as discussed in the ISA, are based not only on
 8    the more expansive epidemiological evidence available in this review but also reflects
 9    consideration of important progress that has been made to advance our understanding of a
10    number of potential biologic modes of actions or pathways for PM-related cardiovascular and
11    respiratory effects (US EPA 2009a, chapter 5).  With respect to suggestive evidence for a
12    broader range of effects, the body of scientific evidence is somewhat expanded but is still limited
13    with respect to associations between long-term PM2 5 exposures and developmental and
14    reproductive effects as well as cancer, mutagenic, and genotoxic effects.  Thus, we reach the
15    preliminary conclusion that there is stronger and more consistent and coherent support for
16    associations between short- and long-term PM2 5 exposure and a broader range of health
17    outcomes than was available in the last review, providing the basis for fine particles at least as
18    protective as the current PM2 5 standards.
19           Having reached this initial conclusion, we then consider how the new evidence informs
20    our understanding of susceptible populations by asking the following question:
21    •  To what extent does the newly available scientific evidence expand our understanding
22       of susceptible populations, including identification of new susceptible populations?
23           As an initial matter, interindividual variation in human responses to air pollutants
24    indicates that some population groups are at increased risk for the detrimental effects of ambient
25    exposure to PM.9 To facilitate the identification of populations at greatest risk for PM-related
26    health effects, studies have evaluated factors that contribute to the susceptibility or vulnerability
27    of an individual to PM. The definition for both of these terms has been found to vary across
28    studies, but in most instances  susceptibility refers to biological or intrinsic factors (e.g., lifestage,
29    gender) while vulnerability refers to non-biological or extrinsic factors (e.g., socioeconomic
30    status [SES]) (see US EPA, 2009a, Table 8-1).  However,  in many cases a factor identified that
31    increases an individual's risk for morbidity or mortality effects from exposure to an air pollutant
32    (e.g., PM) cannot be easily categorized as a susceptibility or vulnerability factor. At times, the
33    terms susceptibility and vulnerability cannot be distinguished from one another. As it has been
      9 Although studies have primarily used exposures to PM10 or PM2 5, the available evidence suggests that the
      identified factors also increase risk from PM10.25 (US EPA, 2009a, section 8.1.8).
            Draft Do Not Quote or Cite           2-25                         March 2010

-------
 1    defined in the ISA (US EPA, 2009a, p. 8-1), the term "susceptible" encompasses both these
 2    concepts. This document maintains that definition.
 3          The 2004 AQCD noted that the existence of heart and lung disease was clearly linked
 4    with increased susceptibility to effects from PM exposure, based on epidemiological and
 5    toxicological studies along with dosimetric evidence. The epidemiological evidence for
 6    increased susceptibility to PM in this population was primarily derived from studies of short-
 7    term exposure. Long-term exposure studies suggested that PM may result in chronic respiratory
 8    disease or decreased lung function growth, thereby increasing susceptibility to short-term
 9    changes in PM.  It was also noted that studies available at that time supported considering older
10    adults and children, including possibly infants, as susceptible groups, recognizing that there is
11    likely overlap between age categories and the preexistence of cardiopulmonary diseases.
12    Emerging evidence indicated that people of lower (SES) or people who have particularly
13    elevated exposures, due to factors such as residential location, may be at greater risk of PM-
14    related effects. The evidence available at that time did not generally allow distinctions to be
15    drawn between the PM indicators, in terms of which groups might be more susceptible to PM2 5
16    and/or PMio-2.5.
17          As discussed below, the ISA concludes that the epidemiological, controlled human
18    exposure and toxicological studies continue to provide evidence of increased risk for various
19    populations, including people at certain life stages, in this case children and older adults, people
20    with pre-existing cardiovascular and respiratory diseases (e.g., asthma), and people with lower
21    SES (US EPA, 2009a, section 2.4.1, chapter 8). This section expands on the discussion in the
22    ISA of susceptible populations in that it also considers supporting evidence for increased risk in
23    these populations from studies that examined only one susceptible population. However it does
24    not contain information from the ISA for potentially susceptible populations for which there was
25    not sufficient evidence to it to draw conclusions about increased susceptibility (e.g., people with
26    diseases that involve chronic inflammation, such as diabetes, or with genetic susceptibility).
27    Children and Older Adults
28          Childhood represents a life stage that has generally been considered susceptible to
29    exposure to PM due to the following factors: children spend more time outdoors; children have
30    greater activity levels than adults; children have exposures resulting in higher doses per body
31    weight and lung surface area; and also because of the potential for irreversible effects  on the
32    developing  lung (US EPA, 2009a, section 8.1.1.2). Recent studies of long-term exposure to
33    PM2.5 have greatly expanded the evidence for effects on lung development in children. The
34    extended follow-up for the Southern California CHS  includes several new studies that report
35    positive associations between long-term exposure to PM2.5 and respiratory morbidity,
36    particularly for such endpoints as lung function growth, respiratory symptoms (i.e., bronchitic
            Draft Do Not Quote or Cite          2-26                        March 2010

-------
 1    symptoms), and respiratory disease incidence (US EPA, 2009a, section 7.3; Gauderman et al.,
 2    2004). New analyses have been conducted that include longer follow-up periods of the CHS
 3    cohort through 18 years of age, provide evidence that effects from exposure to PM2 5 persist into
 4    early adulthood and are more robust and larger in magnitude than reported for shorter follow-up
 5    periods available in 2004. Supporting these results are new cohort studies conducted by other
 6    researchers in other locations with different methods that provide enhanced evidence for
 7    respiratory effects related to long-term exposures to PMi0 that is dominated by PM2.5 (US EPA,
 8    2009a, section 7.3). New studies provide positive associations from Mexico City, Sweden, and a
 9    national cystic fibrosis cohort in the U.S.  A natural experiment in Switzerland reported that
10    improving PM air quality may slow the annual rate of decline in lung function in adulthood,
11    indicating benefits to public health. In addition, investigators of the CHS suggest PM2 5 may also
12    act as a modifier of the association between lung function and asthma, inducing declines in lung
13    function and a concomitant increase in new onset asthma (Islam et al 2007).  Thus, the ISA (US
14    EPA, 2009a, section 7.3.9.1) concludes that the data for respiratory morbidity in children are
15    consistent and coherent across several study designs, locations, and researchers. Preliminary
16    evidence from toxicological studies provides some coherence and biological plausibility for the
17    observed associations with lung function decrements (US EPA, 2009a, section 7.3.2.2).
18          With respect to short-term exposures to PM, the 2004 AQCD found that studies which
19    stratified results by age typically reported associations between short-term exposures to PM and
20    respiratory-related health effects in children with asthma (US EPA, 2004, section 8.4.9). This
21    body of evidence has been strengthened by newly available epidemiological studies that provide
22    evidence of reductions in lung function (FEVi) and  an increase in respiratory symptoms and
23    medication use associated with PM exposure among asthmatic children (US EPA, 2009a,
24    sections 6.3.1  and 6.3.2.1).  These include two large, longitudinal studies in urban areas of the
25    US, as well as a number of smaller panel studies conducted in the U.S. (US EPA, 2009a, section
26    6.3.1).
27          No epidemiologic studies of pulmonary inflammation were described in the 2004 AQCD.
28    Several new panel studies of children, using exhaled NO (eNO) as a biomarker of airway
29    inflammation, found generally positive associations between PM exposure and eNO levels in
30    asthmatic children who did not use inhaled corticosteroids (US EPA, 2009a, section 6.3.3.1).
31          With respect to short-term increases in PM concentrations associated with respiratory -
32    related hospitalizations and ED visits, the epidemiologic evidence presented in the  2004 AQCD
33    was consistent across  studies (US EPA, 2004,  section 8.3.2.5). Newly available evidence
34    provides further support for this relationship, with larger effect estimates observed  among
35    children and older adults. However, effect estimates are clearly heterogeneous and sensitive to

            Draft Do Not Quote or Cite           2-27                        March 2010

-------
 1    choice of lag, with evidence of both regional and seasonal differences (US EPA, 2009a, sections
 2    6.3.8 and 6.3.10).
 3          Two recent toxicological studies, presented in the ISA (US EPA, 2009a, section 8.1.1.2),
 4    provide biological plausibility for the increase in PM-related respiratory effects in children
 5    observed in the epidemiological studies.  These studies include effects on lung function,
 6    pulmonary injury and evidence suggesting that the developing lung is more susceptible to PM.
 7    The ISA (US  EPA, 2009a, p. 8-5) concludes that the evidence from epidemiological studies that
 8    have examined the health effects associated with all size fractions of PM and toxicological
 9    studies that have examined individual PM components provide support for the hypothesis that
10    children are at greater risk of respiratory effects from exposure to PM.
11          With respect to older adults, the 2004 AQCD (US EPA, 2004, section 9.2.4.2) concluded
12    that people aged 65+ years appear to be at somewhat higher risk for PM exacerbation of
13    cardiovascular-related disease effects and, perhaps, tend to experience higher PM-related total
14    (nonaccidental) mortality risk, as well. The higher prevalence of pre-existing cardiovascular and
15    respiratory diseases found in this age range compared to younger age groups increases risk,
16    primarily due to the gradual decline in physiological processes as part of the aging process (US
17    EPA, 2009a, section 8.1.1.1). Therefore, some overlap  exists between this life stage and the
18    group of people with pre-existing diseases.
19          Newly available evidence for PM-related health effects in the older adult life stage spans
20    epidemiologic, controlled human exposure,  and toxicological studies (US EPA, 2009a, section
21    8.1.1.1).  One new large epidemiologic study (US EPA, 2009a, section 7.2.9; Miller et al., 2007)
22    of post-menopausal female residents of 36 U.S. metropolitan areas (age range = 50-79 yr) found
23    increases in PM2.5-related risk of MI, coronary revascularization,10 and their combination with
24    CHD-related  death for participants free of CVD  at baseline.  While recent epidemiological
25    evidence for cardiovascular morbidity effects in  older adults in response to short-term exposure
26    to PM2.s and PMi0-2.5 is limited, when taken together with evidence from studies of PMi0,
27    support is found for increased risk of cardiovascular morbidity in older adults.  With regard to
28    respiratory morbidity, while some epidemiologic studies have reported an increase in respiratory
29    hospital admissions in individuals 65 years of age and older, consistent associations have not
30    been observed across all such studies.
31          With respect to mortality, recent epidemiologic studies have also found that individuals
32    greater than 65 years old are at greater risk of all-cause (non-accidental) mortality upon short-
33    term exposure to both PM2.5 and PMio, consistent with the findings of the 2004 AQCD.
34    Epidemiological studies that examined the association between mortality and long-term exposure
      10 Coronary revascularization includes percutaneous coronary interventions, such as angioplasty.
            Draft Do Not Quote or Cite           2-28                        March 2010

-------
 1    to PM2.5 and stratified the results by age (i.e., less than 65 year of age compared to aged 65 and
 2    older) reported results that are consistent with the short-term exposure studies. However, in
 3    sensitivity analyses studies have found evidence that risk declines with age starting at age 60
 4    until there is no evidence of an association among people age 85 and older.
 5           Controlled human exposure and dosimetry studies provide additional evidence for an
 6    increase in cardiovascular and respiratory effects among older adults. Healthy older subjects
 7    exposed to PM2.5 concentrated ambient particles (CAPs) experienced significant decreases in
 8    HRV (both in time and frequency) immediately following exposure,  when compared to healthy
 9    young subjects and to older adults with COPD (US EPA, 2009a, section 8.1.1.1). Dosimetry
10    studies have shown a depression of fine and coarse PM clearance in all regions of the respiratory
11    tract with increasing age beyond young adulthood.  These results suggest that older adults are at
12    greater risk of PM-related respiratory health effects (US EPA, 2009a, section 8.1.1).
13           Animal toxicological studies have attempted to characterize the relationship between age
14    and PM-related health effects through the development of models that mimic the physiologic
15    conditions associated with older people. For example, arrhythmias have been observed in older,
16    but not younger, rats exposed to PM2.5 CAPs. In addition, studies that used a mouse model of
17    terminal senescence observed various cardiovascular-related responses. Overall, these studies
18    provide biological plausibility for the increase in cardiovascular effects in older adults observed
19    in the controlled human exposure and epidemiologic studies (US EPA, 2009a, section 8.1.1).
20           In summary, the ISA concludes that evidence from epidemiological, controlled human
21    exposure, and toxicological studies provide coherence and biological plausibility for the
22    association between short-term exposure to PM and cardiovascular morbidity in older adults (US
23    EPA, 2009a, section 8.1.1.1).  Additional evidence from epidemiological studies that focus on
24    mortality and respiratory morbidity in response to short-term exposure to PM also indicate that
25    older adults represent a susceptible life stage.
26    Pre-existing Cardiovascular or Respiratory Diseases
27           With regard to the risk from PM exposure to people with pre-existing cardiovascular or
28    respiratory diseases, the 2004 AQCD (US EPA, 2004, section 9.2.4.1) noted that a number of
29    time-series epidemiological studies reported increased risk in study subsets of individuals with
30    preexisting heart or lung diseases.  Toxicological studies using animal models of
31    cardiopulmonary disease provided evidence suggestive of enhanced susceptibility to inhaled PM
32    in "compromised" hosts. Human dosimetry studies of subjects with COPD and asthma indicated
33    that airways disease leads to very heterogeneous distributions of PM deposited within the lung.
34    These studies have shown up to 10-fold higher than normal deposition  at airway bifurcations,
35    thus creating "hot-spots" that may have biological implications.

             Draft Do Not Quote or Cite          2-29                         March 2010

-------
 1           The ISA notes that more recent epidemiological and controlled human exposure studies
 2    have directly examined the effect of PM on individuals with pre-existing diseases and
 3    toxicological studies have employed disease models to identify whether exposure to PM
 4    disproportionately effects certain  populations (US EPA, 2009a, section 8.1.6).
 5           The potential effect of underlying cardiovascular diseases on PM-related health responses
 6    has been examined using epidemiological studies that stratify effect estimates by underlying
 7    conditions or secondary diagnoses, and toxicological studies that use animal models to mimic the
 8    pathophysiological conditions associated with various cardiovascular diseases (e.g., MI,
 9    ischemia, and atherosclerosis).  A limited number of controlled human exposure studies have
10    also examined the potential relationship between cardiovascular diseases and exposure to PM in
11    individuals with underlying cardiovascular conditions, but these studies have provided somewhat
12    inconsistent evidence for these associations.
13           One epidemiological study (US EPA, 2009a, section 7.2.9; Zanobetti and Schwartz,
14    2007) investigated associations between long-term exposure to PMio and the progression of
15    disease or reduced survival in a 21-city study of people discharged following an acute
16    myocardial infarction (MI).  The study found significant associations for mortality, CHF and
17    new hospitalization for MI.  This  is the first long-term study showing a significant association
18    between particle exposure and adverse post-Mi outcomes in people who survived an MI. A few
19    toxicological studies examined potential effect modification of pre-existing cardiovascular
20    conditions on effects of long-term exposure to PM. In studies that focused on the cardiovascular
21    effects following subchronic exposure to PM in ApoE"" mice, pathophysiological effects were
22    observed that included a decreasing trend in heart rate, physical activity, and temperature along
23    with responses in HRV; enhanced size of early atherosclerotic lesions was observed consistently
24    across studies.
25           The majority of the epidemiological literature that examined associations between short-
26    term  exposure to PM  and cardiovascular outcomes focuses on cardiovascular-related hospital
27    admissions (HA) and emergency department (ED) visits.  Hypertension is the pre-existing
28    condition that has been considered to the greatest extent when examining the association
29    between short-term exposure to PM and cardiovascular-related HAs and ED visits.  The results
30    of these studies have been mixed. There is also some new evidence from epidemiological
31    studies that individuals with pre-existing IHD are  at greater risk from PM exposure, with
32    evidence from a toxicological study implicating impaired myocardial blood flow in the response.
33    Additional studies have examined the effects of PM on cardiac function in individuals with
34    dysrhythmia with mixed results. Limited and inconclusive evidence is available from
35    epidemiological studies that examined other pre-existing cardiovascular conditions, such as CHF
36    and MI. Toxicological studies have provided some additional evidence  for the cardiovascular
            Draft Do Not Quote or Cite           2-30                        March 2010

-------
 1    health effects associated with exposure to PM in rodent models of underlying MI, but the
 2    evidence is also inconclusive.  Controlled human exposure studies that examined the effect of
 3    pre-existing diseases on cardiovascular outcomes with exposure to PM are less consistent and
 4    difficult to interpret in the context of the results from the epidemiological and toxicological
 5    studies.
 6          A few new studies have also examined effect modification in mortality associations.
 7    These studies have found evidence of an increase in risk estimates for associations between PMi0
 8    and mortality in individuals with underlying stroke, and effect modification of the PM-mortality
 9    association in individuals with CHF.
10          In summary, the newly available evidence from epidemiological and toxicological, and to
11    a lesser extent, controlled human exposure studies indicates increased susceptibility of
12    individuals with underlying cardiovascular diseases to PM exposure. Although the evidence for
13    some outcomes was inconsistent across epidemiological and toxicological studies, this could be
14    due to a variety of issues including the PM size fraction used in the study, along with the study
15    location.
16          With regard to pre-existing respiratory illnesses, studies evaluated in the ISA have
17    examined the effect of these illnesses on multiple health outcomes (e.g., mortality, asthma
18    symptoms) in response to exposure to ambient concentrations of PM (US EPA, 2009a, section
19    8.1.6.2). Epidemiological studies have examined the effect of short-term exposure to PM on the
20    respiratory health of asthmatic individuals finding an increase in medication use and respiratory
21    symptoms (i.e., asthma symptoms, cough, shortness of breath, and chest tightness) with short-
22    term exposure to PM2 5, and morning symptoms and asthma attacks with short-term exposure to
23    PMio. Toxicological  studies provide evidence that PM exposure results in allergic sensitization,
24    and that individuals with allergic airways conditions are at greater risk of allergic airways
25    responses upon exposure to PM2.5. Further, there is much more  limited evidence which suggests
26    that non-allergic respiratory morbidities may also increase the susceptibility of an individual  to
27    PM-related respiratory effects.
28          The results from the epidemiological and toxicological studies that focused on underlying
29    allergic airways disease are supported by a series of controlled human exposure studies which
30    have shown that exposure to diesel exhaust particles (DEPs) increases the allergic inflammatory
31    response in atopic individuals. However, other studies reported  that healthy and asthmatic
32    subjects exposed to coarse, fine and ultrafme CAPs, exhibited similar respiratory responses,
33    although these  studies excluded moderate and severe asthmatics that would be expected to show
34    increased susceptibility to PM exposure.
35          With respect of pre-existing COPD, the results of epidemiological studies have been
36    mixed. A few  controlled human exposure studies examined the  effects of exposure to PM2.5
            Draft Do Not Quote or Cite           2-31                         March 2010

-------
 1    CAPs on healthy and COPD subjects and found no significant difference in respiratory effects.
 2    However, the results from dosimetric studies have shown that COPD patients have increased
 3    dose rates and impaired mucociliary clearance relative to age-matched healthy subjects,
 4    suggesting that individuals with COPD are potentially at a greater risk of PM-related health
 5    effects (US EPA, 2009a, section 3.2.1).
 6          A few epidemiological studies examined the effect of underlying respiratory illnesses on
 7    the association between short- and long-term exposure to PM and mortality. Using different
 8    pre-existing respiratory illnesses, two studies found that short-term exposure to PMio increased
 9    the risk of non-accidental mortality in individuals with a secondary diagnosis of pneumonia and
10    circulatory mortality in individuals with a secondary diagnosis of a respiratory illnesses. Another
11    study observed an association between long-term exposure to PMio and mortality in individuals
12    that had previously been hospitalized for COPD.  Together, these studies show the potential
13    effect of underlying respiratory illnesses on the PM-mortality relationship (US EPA, 2009a, p. 8-
14    13).
15          In summary, with respect to the potential for increased risk from PM exposure to people
16    with respiratory disease, overall, the controlled human exposure and toxicological studies
17    evaluated in the ISA provide biological plausibility for the increased risk of health effects
18    observed in epidemiological studies among asthmatic  individuals in response to PM exposure
19    (US EPA, 2009a, section 8.1.6.2). The evidence from studies that examined associations
20    between PM and health effects in individuals with COPD is inconsistent.
21    Socioeconomic Status
22          Socioeconomic status is a composite measure that usually consists of economic status,
23    measured by income; social status measured by education; and work status measured by
24    occupation. Based on data from the U.S. Census Bureau in 2006, from among commonly-used
25    indicators of SES, about 12% of individuals and 11%  of families are below the poverty line (US
26    EPA, 2009a, section 8.1.7). Within the U.S. approximately 16% of the population does not have
27    a high school degree and only 27% have a bachelor's  degree or higher level of education (US
28    Census,  2009). Educational attainment generally coincides with  an individual's income level,
29    which is correlated to other surrogates of SES, such as residential environment. Additionally,
30    low SES individuals have been found to have a higher prevalence of pre-existing diseases;
31    limited access to medical treatment; and limited access to fresh foods leading to a reduced intake
32    of antioxidants, polyunsaturated fatty acids and vitamins, which can increase this population
33    group's risk from PM.  Low SES and surrogates of SES such as educational attainment,
34    residential location and nutritional status have been shown in some studies to modify health
35    outcomes of PM exposure for a population.

            Draft Do Not Quote or Cite           2-32                        March 2010

-------
 1          The 2004 AQCD concluded there is effect modification of long-term PM exposures-
 2    mortality associations due to socioeconomic factors. In the ACS and Harvard Six Cities cohort
 3    analyses on mortality risk with long-term exposure to PM2 5, there was clear evidence of greater
 4    effects being reported in the cohort subgroups with lower education levels (US EPA 2004,
 5    section 9.2.4.5).  The ISA concludes that this is further supported by a reanalysis of the ACS
 6    cohort (Krewski et al., 2009), which found moderate evidence for increased lung cancer
 7    mortality in individuals with a high school education or less in response to long-term exposure to
 8    PM2.5 (US EPA, 2009a, section 8.1.7). Another study examined whether long-term exposure to
 9    traffic-related pollutants varied by SES at the block group level.  The authors found higher
10    concentrations of NC>2 associated with lower SES areas, which suggests that lower SES
11    individuals are disproportionately exposed to traffic-related pollutants, including PM.
12          Among the studies of short-term  PM exposure, the 2004 CD concluded that the evidence
13    was mixed regarding SES and PM-related health risks. New studies evaluated in the ISA
14    indicate that evidence of the influence of SES on health outcomes related to short-term exposures
15    is stronger (US EPA, 2009a, section 8.1.7). These include an increased risk in mortality
16    associated with short-term exposure to PM2.5 and its components. Jerrett  et  al. (2004) examined
17    the modification of short-term mortality  effects due to particulate air pollution exposure by
18    residential location in Hamilton,  Canada. The authors found that the area of the city with the
19    highest SES characteristics (measured using the surrogate educational attainment) displayed no
20    evidence of effect modification while the area with the lowest SES characteristics had the largest
21    health effects. Another study noted an increased risk in mortality associated with short-term
22    exposure to PM2 5 and its components for individuals with low SES, while additional analyses
23    stratified by education level have also observed consistent trends of increased mortality for PM2.5
24    and PM2.5 species for individuals with low educational attainment (US EPA 2009a, section
25    8.1.7).
26           Nutritional deficiencies have been associated with increased susceptibility to a variety  of
27    infectious diseases and chronic health effects. Low SES may decrease access to fresh food,
28    thereby nutritional deficiencies could increase susceptibility to PM-related health effects. One
29    study that examined the association between exposure to PM2.5 and HRV  in  individuals with
30    genetic polymorphisms associated with increased risk of CVD, found that when individuals with
31    these genetic polymorphisms increased their intake of B6, B12, or methionine, no PM2.5 effect  on
32    HRV was observed.
33          The ISA concludes that there is evidence that SES, measured using surrogates such as
34    educational attainment or residential location, modifies the association between PM and
35    morbidity and mortality outcomes (US EPA, 2009a, section 8.1.8). In addition, nutritional

            Draft Do Not Quote or Cite           2-33                        March 2010

-------
 1    status, another surrogate of SES, has been shown to have protective effects against PM exposure
 2    in individuals that have a higher intake of some vitamins and nutrients.
 3    Summary
 4          In summary, we reach the preliminary conclusion that there are several population groups
 5    that are likely to be susceptible to PM-related effects.  These groups include the life stages of
 6    children and older adults, those with preexisting heart and lung diseases, and those of lower SES.
 7    We also preliminarily conclude that the available evidence does not generally allow distinctions
 8    to be drawn between the PM indicators, in terms of which groups might have greater
 9    susceptibility to PM2.5 and PMio-2.5-
10    •  To what extent does the newly available scientific evidence report associations that
11       extend to air quality concentrations that are lower than had previously been observed
12       or that are observed in areas that would likely meet the current suite of PMi.s
13       standards?
14          We now focus our attention on addressing the question of whether the available evidence
15    supports consideration of standards that are more protective than the current suite of PM2.5
16    standards.  In addressing this question, we first recognize that the ISA concludes there is no
17    evidence to support the existence of a discernable threshold below which effects would not occur
18    (US EPA, 2009a, section 2.4.3). We consider whether new evidence provides information for
19    health effects associated with air quality levels that are lower than had previously been observed,
20    in particular to levels at or below the level of the current annual standard (15 |ig/m3) and/or to
21    levels at or below the level of the current 24-hour standard (35 jig/m3).  As an initial matter, we
22    first focus on levels at which effects classified as having a causal or likely causal association
23    with PM2.5 exposure have been observed.  We consider effects classified as having suggestive
24    associations with PM2.5 exposure in section 2.3 when considering the margin of safety provided
25    by alternative suites of standards. We have evaluated the air quality data using the forms of the
26    current PM2.5 standards.n
27    Associations with Long-term PM; s Exposure
28          With regard to mortality associated with long-term PM2.5 exposure, in the last review, the
29    2004 AQCD placed greatest weight on the reanalyses and extensions of two prospective cohort
30    studies, the ACS and the Harvard Six Cities studies, finding that these studies provided "strong
31    evidence" for associations  with fine particles and confirmed and strengthened the evidence
32    available in the review completed in 1997 (US EPA, 2004, p. 9-33; Krewski et al., 2000). In the
      11 The current form of the annual PM25 standard is the annual arithmetic mean, averaged over 3 years with limited
      conditions allowing spatial averaging.  The current form of the 24-hour PM25 standard is the 98th percentile,
      averaged over 3 years.
             Draft Do Not Quote or Cite           2-34                        March 2010

-------
 1    current review, as stated above, staff notes that newly available extended follow-up analyses of
 2    the ACS and Harvard Six Cities studies provide consistent and stronger evidence of a causal
 3    association with mortality at lower air quality distributions of PM2 5.
 4           In considering studies of the ACS cohort, in the last review, the original and reanalysis
 5    studies reported significant associations between fine particles and mortality (US EPA, 2004,
 6    8.2.3.2.1; Pope et al., 1995; Krewski et al., 2000).  In addition, an extended analysis using the
 7    ACS cohort also available in the last review doubled the original follow-up period to more than
 8    16 years. This study continued to report statistically  significant associations with long-term
 9    PM2.5 exposure with the inclusion of more recent PM2.5 air quality data and triple the number of
10    deaths (US EPA, 2004, 8.2.3.2.2; Pope et al., 2002).  As with the original ACS cohort study, no
11    evidence of a threshold was observed in the  relationships with total, cardiovascular-related, and
12    lung cancer mortality reported in this extended study. A recent extended analysis of the ACS
13    cohort available in this review lengthens the follow-up of this important study to 18 years (1982
14    to 2000) (US EPA, 2009a, pp. 7-84 to 7-85;  Figure 7-6; Krewski et al., 2009). This extended
15    analysis has "confirmed with remarkable consistency the association of mortality and exposure
16    to PM2 5 reported in previous studies of the ACS data starting 15 years ago... and added
17    precision, especially for the evidence that ischemic heart disease is  a cause of death particularly
18    affected by exposure" to PM2.5 (HEI, 2009, p.  135). We note the decline in ambient PM2.5
19    concentrations over 18 years, from an aggregate long-term mean in 58 metropolitan statistical
20    areas (MSAs) of 21.2 (4.6)12 |ig/m3 (ranging from approximately 9 to 34 |ig/m3 across cities) in
21    the original monitoring period (1979 to 1983) declining to air quality distributions reported for
22    the most recent years evaluated (1999-2000) for 116  MSAs with an aggregate long-term mean
23    PM2.5 concentration of 14.0 (3.0) |ig/m3 (ranging from approximately 6 to 22 |ig/m3 across cities)
24    to an overall average across all year of 17.1  (3.7) |ig/m3 (ranging from about 7.5 to 30
25    |ig/m3)(US EPA, 2009a, Figure 7-6; Krewski et al, 2009; Pope et. al, 2004). In addition, Eftim et
26    al., (2008) assessed the ACS sites using an ecological cross-sectional study design of a Medicare
27    cohort, incorporating more recent air quality (2000-2002) with an aggregate long-term mean
28    PM2.5 concentration of 13.6 (2.8) |ig/m3 (ranging from approximately 6 to 25 |ig/m3).  This study
29    reported somewhat higher effect estimates than those reported by the original investigators. The
30    ISA concludes there may be several possible explanations for this apparent increase, especially
31    that this is an older population or more likely because of the lack of personal confounder
32    information (e.g., past personal smoking information) that "led to an insufficient control for the
33    effects of these other variables' effects on mortality, inflating the pollution effect estimates
      12 Throughout this discussion concentrations reported are mean (standard deviation, SD) or median (interquartile
      range, IQR, 25 to 75%).
            Draft Do Not Quote or Cite           2-35                        March 2010

-------
 1    somewhat, similar to what has been found in the ACS analyses when only ecological-level
 2    control variables were included" (US EPA, 2009a, pp. 7-87 to 7-88; Figures 7-6 and 7-7).
 3           With respect to a the Harvard Six Cities study, Laden et al (2006) have extended
 4    mortality follow-up for an additional eight years during a period of "reduced air pollution
 5    concentrations" lengthening the follow-up period to 25 years.  These investigators reported
 6    statistically significant associations between long-term exposure to PM2.5 and mortality for air
 7    quality data for the original study period (1974 to 1989) as well as the extended follow-up period
 8    (1990-1998). As noted in the ISA, this study provides evidence of a statistically significant
 9    reduction in mortality risk with decreasing long-term PM2.5 concentrations (US EPA, 2009a, p.
10    7-84).  Evidence from the original and reanalysis of the Harvard Six Cities study available in the
11    last review reported an aggregate long-term mean PM2 5 concentration of 18 |ig/m3 (ranging from
12    approximately 11 to 30 |ig/m3 across cities) (US EPA, 2004, p. x.x; Dockery et al., 1993;
13    Krewski et al., 2000). In the extended follow-up period, an aggregate long-term mean PM2 5
14    concentration of 14.8 (4.2) |ig/m3 (ranging from approximately 10 to 22 |ig/m3 across cities) with
15    an overall aggregate mean across both study periods of 16.4 (5.6) |ig/m3 (ranging from
16    approximately 11 to 25 |ig/m3) (US EPA, 2009a, Figure 7-6; Laden et al., 2006; Laden, 2009).
17    This reduction was observed for total mortality as well as cardiovascular-related and respiratory-
18    related mortality but not deaths related to lung cancer, "a disease with a longer latency period
19    and less reversibility" (Laden et al, 2006). The ISA notes that "a statistically significant
20    reduction in mortality risk [was] reported with reduced long-term PM2.5 concentrations" (US
21    EPA, 2009a, p. 7-84). These findings suggest that the mortality effects of long-term air pollution
22    may be at least partially reversible over periods of a decade" (Laden et al., 2006; Table 3). In an
23    additional analysis of the extended follow-up of the Harvard Six Cities cohort study,
24    investigators reported the  concentration-response relationship was linear and "clearly continuing
25    below the level" of the current annual standard (US EPA, 2009a, p. 7-92; Schwartz et al., 2008).
26    Eftim et al. (2008) also analyzed the Harvard Six Cities study sites using  a Medicare cohort with
27    more recent air quality data (2000-2002) and reported consistent but somewhat higher effect
28    estimates with relation to the extended follow-up of this study conducted by Laden et al. (2006)
29    (US EPA, 2009a, pp. 77-87-7-88, Figures 7-6 and 7-7).  The aggregate long-term mean reported
30    in this study was 14.1(3.1) |ig/m3 (ranging from about 10 to 19 |ig/m3) (Eftim et al., 2008).
31           In addition to extended follow-up analyses of the ACS and Harvard  Six Cities studies,
32    new evidence is available in this review from two new U.S. long-term cohort studies. The
33    Women's Health Initiative (WHI) provides evidence of association between long-term PM2 5
34    exposure and cardiovascular morbidity  and mortality effects. This study reported results
35    consistent with the ACS and Harvard Six Cities studies while identifying much larger relative
36    risk estimates per |ig/m3.  This study represents an important new cohort of postmenopausal
             Draft Do Not Quote or Cite          2-36                        March  2010

-------
 1    women with no previous history of pre-existing cardiac disease, potentially a "healthier cohort
 2    population" than that considered by the ACS and Harvard Six Cities studies (US EPA, 2009a, p.
 3    7-87).  The ISA notes, that the "PM2 5 impacts may yield higher relative risk estimates in the
 4    WHI population because the PM2.5 risk is being compared to a much lower prevailing risk of
 5    cardiovascular death in this select study population" (US EPA, 2009a, p. 7-87).  The overall
 6    PM2.5 concentration averaged across cities reported in this study was 13.5(3.3) |ig/m3 (ranging
 7    from about 3.4 to 28 |ig/m3) (Miller et al., 2007).
 8           A new retrospective cohort study, the Medicare Cohort Air Pollution Study (MCAPS)
 9    has evaluated long-term PM2.5 exposure and mortality risk in older adults (65 years of age and
10    older) within 3 geographic regions. Zeger et al. (2007) reported that long-term exposure to PM2 5
11    was significantly associated with an increase in mortality among Medicare participants, and
12    stronger associations were observed in eastern counties compared to a national estimate, and  no
13    association was observed among Western counties. However, effect estimates decreased by 50%
14    with adjustment for spatial confounding.  In a subsequent retrospective cohort study  (MCAPS),
15    Zeger et al. (2008), reported that average 6-year exposure to PM2.5 was significantly  associated
16    with increased risk of mortality in the eastern and central regions.13 In addition, the results
17    indicated that risk declined with increasing age, and similar to their earlier study, no  association
18    was observed between PM2 5 and mortality in the western region.  Moreover, risk estimates were
19    similar to effect estimates  reported in the ACS and Harvard Six Cities study, providing
20    coherence across prospective studies  for an association between long-term exposure  to PM2.5 and
21    mortality.  This study reported associations between long-term PM2 5 exposure and mortality  for
22    the eastern and central regions that were qualitatively similar to those reported in the ACS and
23    Harvard Six Cities studies. The long-term aggregate PM2.5 median concentration reported
24    across all cities was 13.2 |ig/m3 (with an interquartile range from about 11 to 15 |ig/m3) (US
25    EPA, 2009a, p. 7-88; Zeger et al., 2008).14
26           With respect to respiratory morbidity effects associated with long-term PM2.5 exposure,
27    the 2004 AQCD concluded that new studies of a cohort of children in  Southern California had
28    built upon earlier limited evidence to provide evidence that long-term  exposure to fine particles
29    was associated with development of chronic respiratory disease and reduced lung function
30    growth (US EPA, 2004, p. 9-33). The across-city mean of 2-week average PM25 concentrations
31    reported in the initial Southern California Children's Health Study was approximately 15.1
      13 The Zeger et al. (2008) analysis included COPD as a proxy for smoking status.  The investigators reported the risk
      estimate for the eastern region declined and the central region increased using this adjustment. This result may
      possibly be related to bias introduced in using COPD as a proxy for smoking.
      14 Zeger et al. (2008) assessed the relative risk of death associated with long-term PM2 5 exposure for three U.S.
      regions. Median (IQR) PM2 5 concentrations reported by region were: Eastern region -14.0 ug/m3 (12.3-15.3);
      Central region- 10.7 ug/m3 (9.8-12.2); Western region - 13.1 ug/m3 (10.4-18.5).
             Draft Do Not Quote or Cite           2-37                        March 2010

-------
 1    |ig/m3 (ranging from approximately 7 to 32 |ig/m3) (Peters et al., 1999). These results were found
 2    to be consistent with results of cross-sectional analyses of the 24-City study by Dockery et al.
 3    (1996) and Raizenne et al. (1996) which reported a long-term cross-city mean PM2 5
 4    concentration of 14.5(4.2) |ig/m3 (ranging from approximately 6 to 21 |ig/m3). Gauderman et al.
 5    (2004) have now extended the analysis of the Southern California Children's Health Study to 8
 6    years, following the children between the ages of 10 and 18, "a period of rapid lung
 7    development" and reported that the "pollution-related deficits in the average growth in lung
 8    function over the 8-year period resulted in clinically important deficits in attained lung function
 9    at the age of 18" (US EPA, 2009a, p. 7-27).  The four year cross-city mean was 13.8 (7.7) |ig/m3
10    (ranging from approximately 6 to 29 |ig/m3) across the 12 study communities (McConnell et al.,
11    2003).
12          In summary, newly available scientific evidence provides support for associations
13    between long-term PM2.5 exposure and mortality and morbidity effects that extend to air quality
14    levels that are lower than had previously been observed.  These studies evaluate a broader range
15    of health outcomes than were considered in the last review and include extended follow-up for
16    prospective epidemiological studies that were important in the last review as well as additional
17    evidence in important new cohorts. In looking across the body of scientific evidence and
18    focusing on the most recent years of air quality data considered in these studies, we observe
19    effects occurring below the level of the current annual PM2 5 standard (US EPA, 2009a, chapter
20    7; Figures 2-2 and 7-7).
21    Associations with Short-term PMi.s Exposure
22          In the last review, in selecting the level of the 24-hour PM2 5 standard, emphasis was
23    placed on short-term exposure studies and their 98th percentile air quality values that provided
24    evidence of associated health effects in areas that would have met the then current annual and
25    24-hour PM2.5 standards during the time of the study. In focusing on the 98th percentile value in
26    these studies the Administrator recognized that these studies did not provide evidence of clear
27    effect thresholds or lowest-observed levels.  Therefore, the Administrator  sought to establish a
28    standard level that would require improvements in air quality generally in areas in which the
29    distribution of daily short-term exposure to PM2 5 can reasonably be expected to be associated
30    with serious health outcomes (i.e., mortality, respiratory and cardiovascular morbidity). In the
31    prior review, we noted an overall  pattern of statistically significant associations observed in
32    studies of short-term exposure to PM2 5 across a wide range of 24-hour average 98th percentile
33    values in predominately single-city studies as well as a limited number of multi-city studies.  A
34    predominance of studies with 98th percentile values down to approximately 39 |ig/m3 (Burnett
35    and Goldberg, 2003) reported statistically significant associations with mortality, hospital

            Draft Do Not Quote or Cite          2-38                         March 2010

-------
                                                                               th
 1    admissions, and respiratory symptoms.  Within the range of 24-hour average 98  percentile
                                               3
 2    PM2.5 concentrations of about 35 to 30 |ig/m , we did not observe a preponderance of statistically
 3    significant results.  Furthermore, the limited number of studies in which the 98th percentile values
 4    were below this range did not provide a basis for reaching conclusions about associations at such
 5    levels (71 FR 61168/3 - 61169/2, October 17, 2006).
 6           As discussed in section 2.1.3, our current approach to reaching preliminary conclusions
 7    about the adequacy of the current suite of PM2.5 standards is more integrative than the approach
 8    used in the last review. In this review, in considering the adequacy of the current suite of
 9    standards, in evaluating the currently available evidence from short-term PM2.5  exposure studies,
10    we look both at the aggregate long-term mean PM2.5 concentration in key studies as well as the
11    distributions of 24-hour PM2 5 concentrations, with a focus on the 98th percentile concentrations
12    to match the form of the current 24-hour standard, to the extent such data are available.
13           A number of new multi-city and single-city short-term PM2.5 exposure studies and
14    additional analysis of a previously existing cohort study available in this review examine the
15    association between short-term exposure to PM2.5 and mortality and a broader range of
16    cardiovascular and respiratory morbidity endpoints. Multi-city studies support a largely positive
17    and frequently statistically significant relationship between short-term exposure to PM2.5 and
18    increased risk of mortality.  In a multi-city time-series analysis of 112 U.S. cities, Zanobetti and
19    Schwartz (2009) reported that an overall 24-hour average15 PM25 level across all years of 34.3
20    (8.8) |ig/m3 (ranging from 17.9 to 80.3 |ig/m3) (US EPA, 2009a, Figure 6-24) from lag 0-1 days
21    were positively and significantly associated with all-cause, cardiovascular-related (e.g.,
22    myocardial infarction, stroke), and respiratory-related mortality. Furthermore,  city-specific
23    effect estimates included in Figure 6-24 of the ISA indicate the association between short-term
24    exposure to PM2 5 and total mortality and cardiovascular mortality and respiratory mortality
25    endpoints is consistently positive for an overwhelming majority (99%) of the 112 cities across a
26    wide range of air quality concentrations (US  EPA, 2009a, p. 6-178 to  179).  In addition, the
27    authors report that  for all-cause mortality city-specific effect estimates were statistically
28    significant for 55% of the 112  cities, with 24-hour PM2.5 levels ranging from 18.4 to 64.9 |ig/m3.
29           In the current review of evidence on mortality associated with short-term exposure to
30    PM2 5, we note that an expansion of the multi-city Canadian study from 8 (Burnett and Goldberg,
31    2003) to 12 Cities (Burnett et al 2004) yielded results consistent with prior findings of a positive
32    and statistically significant association between an average 1-day lag 24-hour PM2 5 exposure
33    across all years of 38.1 |ig/m3(extending from 25 to 50 |ig/m3) and daily mortality (US EPA,
34    2009a, Figure 2-1). However, the influence of NO2 and limited PM2 5 data for several years
       1 Concentrations reported for short-term studies are mean 98th percentile (standard deviation, SD)
             Draft Do Not Quote or Cite           2-39                         March 2010

-------
 1    during the study period somewhat diminish these findings and underscore the need for more
 2    frequent PM2.5 measurements and additional data on co-pollutants to inform our understanding of
 3    the PM2 5-mortality association.  In general, as can be seen in Figure 6-27 of the ISA, effect
 4    estimates for associations between mortality and short-term exposure to PM2.5 are positive and a
 5    number are statistically significant (ISA, US EPA, 2009a, p. 6-184). Larger effect estimates
 6    were reported in some studies for associations with respiratory mortality in comparison to
 7    cardiovascular and all-cause mortality, but these effect estimate include larger confidence
 8    intervals (i.e. less precision) since respiratory deaths comprise only  a small proportion of total
 9    deaths.
10           With regard to cardiovascular and respiratory morbidity effects, in the first analysis of the
11    MCAPS cohort conducted by Dominici et al (2006a) across 204 US counties, investigators
12    reported a statistically significant association between an average 24-hour PM2.5 concentration
13    across all years of 34.8 |ig/m3 (extending from 10.4 to 85.6 |ig/m3) and hospitalizations for
14    cardiovascular and respiratory diseases. Furthermore, a sub-analysis restricted to days with 24-
15    hour average concentrations of PM2.5 at or below 35 |ig/m3 indicated that, in spite of a reduced
16    statistical power from a smaller number of study days, statistically significant associations were
17    still observed between short-term exposure to PM2.5 and hospital admissions for cardiovascular
18    and respiratory diseases (Dominici, 2006b16)). These results along with the observation that
19    approximately 50% of PM2 5 levels across the 204 counties across all years were below the 24-
20    hour standard suggests that the overall health effects observed across the U.S. are not primarily
21    driven by the higher end of the PM2 5 air quality distribution (Cite Docket # for AQ provided by
22    authors). In an extended analysis of the MCAPS study, Bell et al (2008) reported a positive and
23    statistically significant increase in cardiovascular hospitalizations associated with  and averageO-
24    day lag PM2 5 concentration across all years of 34.2 (8.5) |ig/m3 (extending from 9.4 to 77.0
25    Hg/m3). In addition, significant increases in respiratory disease hospitalizations were associated
26    with mean PM2.5 at lag 2. Furthermore, several single-city short-term PM2.5 exposure studies
27    with average 98th percentile PM2 5 exposures below 35 |ig/m3substantiate findings from the
28    aforementioned multi-city studies generally reporting positive statistically significant and non-
29    significant associations for  PM2 5 exposures in relation to mortality, cardiovascular and
30    respiratory hospitalizations and ED visits (Figure 2-4).  Consequently, these short-term studies
31    provide evidence of PM2.5 associated mortality and morbidity effects occurring at and below the
32    level of the current 24-hour standard.
      16 This sub-analysis was not included in the original publication (Dominici et al., 2006a). Authors provided sub-
      analysis results for the Administrator's consideration as a letter to the docket following publication of the proposed
      rule in January 2006.
             Draft Do Not Quote or Cite           2-40                         March 2010

-------
 1           In considering long-term average ambient concentrations from short-term PM2.5 exposure
 2    studies, we recognize that in the last review, several U.S. and Canadian studies provided
 3    evidence of associations between short-term exposure to PM2 5 and serious health effects in areas
 4    with PM2.5 levels at and above the level of the current annual PM2.5 standard (15 |ig/m3).
 5    Moreover, a few short-term PM2 5 exposure studies (Figure 2-4) available in that review provided
 6    evidence of statistically significant associations with PM2.5 in relation to cardiovascular and
 7    respiratory effects for areas in which long-term average PM2.5 concentrations ranged between 12
 8    and 14 |ig/m3 and 98th percentile 24-hour concentrations ranged between 31 and 59 |ig/m3 (US
 9    EPA, 2005, p. 5-7).
10           Similarly, in this review, many short-term PM2 5 exposure studies, and in particular those
11    conducted across multiple cities, report statistically significant effects with mean concentrations
12    below the level of the current annual standard. Dominici et al. (2006a) and Bell et al. (2008)
13    report 24-hour average PM2 5 levels across all years of 13.4 (2.9) |ig/m3 (extending from 4.4 to
14    22.7 |ig/m3) and 12.9 (2.7) |ig/m3(extending from 4.3 to 20.4 |ig/m3), respectively (US EPA,
15    2009a, Figure 2-1).  Likewise,  Zanobetti and Schwartz (2009) report city annual averages across
16    all years of 13.2 () |ig/m3 (extending from 6.6 |ig/m3 to 24.7  |ig/m3) (Figures 2-1 and 6-24, US
17    EPA, 2009a). In addition, Burnett et al. 2004, reports city annual averages across  all years of
18    12.8 |ig/m3 (extending 8.1 to 16.7 |ig/m3) (Figure 2-1, ISA - US EPA, 2009a). Based on the
19    findings from these multi-city short-term PM2 5 exposure studies, the ISA concludes that overall,
20    consistent positive associations have been reported for a range of mortality and cardiovascular
21    and respiratory morbidity effects for mean 24-hour PM2 5 concentrations at and above 12.8 |ig/m3
22    (US EPA, section 2.3.1.1).
23           Taken together,  these findings from single- and multi-city epidemiological  studies of
24    short-term PM2 5 exposure in relation to mortality and cardiovascular and respiratory morbidity
25    provide evidence of short-term  PM2.5 associated health effects occurring at or below the current
26    levels of the 24-hour  standard (Figure 2-1, US EPA, 2009a, p. 2-14).  These findings are
27    bolstered by evidence of statistically significant PM2.s associated health effects occurring in
28    analyses restricted to days in which 24-hour average PM2.5 levels were below 35 |ig/m3
29    (Dominici, 2006b). Multi-city  short-term studies conducted  since the prior review (Burnett et al.
30    2004; Dominici et al. 2006a; Bell et al. 2008; Zanobetti and Schwartz, 2009) also provide
31    supportive evidence for short-term PM2.5 exposure associated health effects occurring at levels at
32    and below the annual standard.
33    Summary
34           In evaluating the currently available scientific evidence, we reach the preliminary
35    conclusion that the evidence from long and short-term PM2 5 exposure studies, and in particular
36    those studies conducted across multiple U.S. cities calls into question whether the current suite of
             Draft Do Not Quote or Cite           2-41                        March 2010

-------
 1    PM2.5 primary standards protects public health with an adequate margin of safety from effects
 2    associated with long- and short-term exposures to PM2.5.  We also reach the preliminary
 3    conclusion that this evidence provides strong support for considering fine particle standards that
 4    would impart increased protection beyond that afforded by the current annual and 24-hour PM2.5
 5    standards. More protective standards would reflect the substantially stronger and broader body
 6    of evidence for mortality and cardiovascular and respiratory morbidity effects now available in
 7    this review both at lower levels of air quality than had previously been observed and at levels
 8    below the current annual (15 |ig/m3) and 24-hr (35 |ig/m3) PM2.5 standards.

 9    2.2.2   Risk-based Considerations
10           Looking beyond evidence-based considerations, staff also has considered the extent to
11    which health risks estimated to occur upon just meeting the current suite of PM2.5 standards may
12    be judged to be important from a public health perspective, taking into account key uncertainties
13    associated with the estimated risks.
14           The quantitative risk assessment conducted in this review (US EPA, 2010a) builds  upon
15    and expands upon the analyses conducted for the previous review (US EPA, 2005, chapter 4;
16    Abt, 2005). As an initial matter, we recognize that the previous quantitative risk assessment
17    incorporated alternative assumed cutpoints as surrogates for potential population thresholds.  In
18    the current review, the ISA examined the available epidemiologic evidence to characterize the
19    shape of the concentration-response (C-R) relationship and assess possible PM "thresholds"  (i.e.,
20    levels which PM2 5 concentrations must exceed in order to elicit a health response).  Overall,
21    based on limited evidence primarily focusing on cardiovascular-related hospital admissions and
22    emergency department visits associated with short-term exposures to PMio and mortality
23    associated with long-term exposure to PM2 5, using a variety of methods and models, the ISA
24    concludes the currently available evidence supports the use of a no-threshold, log-linear model
25    (US EPA, 2009a, p. 2-25).  Therefore, the quantitative risk assessment conducted for this review
26    did not include assumed cutpoints as surrogates for potential population thresholds.  We
27    concluded that is was more appropriate to focus on alternative rollback strategies.17
28           The assessment of uncertainty and variability completed for this analysis is more
29    comprehensive than had been done for previous risk assessments. This reflects, in part, the
30    development of methods by EPA  staff to  address potentially important sources of variability and
      17 While we have not included cutpoints in the risk assessment for the reasons cited, in the case of long-term
      exposure-related mortality, we have limited estimation of risk to ambient PM2 5 levels above the lowest measured
      level (LML) of the epidemiological study from which the concentration-response function was obtained (Krewski et
      al., 2009; US EPA, 2010a, section 3.2.3). In the case of short-term exposure-related mortality and morbidity, we
      modeled risk down to PRB. In both cases, this reflects consideration for the range of ambient PM2 5 where we
      believe we have confidence in characterizing the nature of the CR function shape (and not an assumption that risk is
      negligible below these levels).
             Draft Do Not Quote or Cite          2-42                         March 2010

-------
 1    uncertainty. For example, to more fully explore potential variability in the patterns of reductions
 2    in ambient PM2.5 that may occur upon just meeting the current and alternative standards, we
 3    incorporated as part of the sensitivity analysis, two additional rollback approaches (hybrid and
 4    peak shaving) in addition to the proportional rollback used in the core analysis. In addition,
 5    recently published literature has allowed us to more rigorously examine the impact of uncertainty
 6    related to specifying C-R functions for long-term exposure-related mortality through a series of
 7    sensitivity analyses (i.e., Krewski et al. (2009) provided extensive analysis of alternative model
 8    specifications for mortality which could be readily incorporated into our sensitivity analysis).
 9           For this review, we have estimated risk for a set of health effects endpoints that reflected
10    consideration of the degree of support in the literature for a causal relationship between PM2.5
11    exposure and the health effect of interest as assessed in the ISA, together with consideration of
12    the health significance of the endpoint.18 Specifically, we have estimated risks for (a) all-cause,
13    IHD, cardiopulmonary and lung cancer mortality related to long-term PM2.5  exposure, (b) non-
14    accidental, cardiovascular (CV) and respiratory mortality related to short-term PM2.5 exposure,
15    and (c) cardiovascular-related and respiratory-related hospital admissions (HA) and asthma-
16    related emergency department visits associated with short-term PM2.5 exposure. While we have
17    modeled risk for a selection of long-term and short-term  exposure-related endpoints, in the
18    discussion of risk estimates presented below, we focus on cardiovascular-related endpoints, since
19    the causal association for these endpoints based on available literature is assessed in the ISA to
20    be the strongest of the endpoints considered.
21           In considering the health risks estimated in selected urban study areas to occur upon just
22    meeting the current suite of PM25 standards, we focused  on a core (primary) set of risk results
23    based on the application of modeling element choices (e.g., concentration-response functions, lag
24    periods) that have the greatest overall support in the literature (hereafter referred to as the "core"
25    results).  In addition, to gain insights into which sources of uncertainty may have the greatest
26    impact on risk estimates when acting alone, or in combination with other sources of uncertainty,
27    as noted above, we conducted a series of single-element and multi-element sensitivity analyses to
28    generate a broader set of reasonable alternative risk estimates that allowed us to place the results
29    of the core analysis in context with regard to uncertainty.  We also conducted additional analyses
30    to place the results of the urban study area analysis into a broader context for characterizing
31    potential national risks. While the core risk estimates receive primary focus in the discussion of
32    risk estimates presented below,  we do references the results of these  additional  supplemental
      18 In addition, we considered whether sufficient information existed in the literature to develop C-R functions and
      whether we could obtain baseline incidence data necessary to generate risk estimates with a reasonable degree of
      confidence.
             Draft Do Not Quote or Cite           2-43                         March 2010

-------
 1    analyses in addressing overall confidence associated with the core risk estimates and in placing
 2    the core risk estimates in a broader national-context.
 3           An important factor to consider in interpreting risk results is that the magnitude of both
 4    long- and short-term exposure-related risk depends primarily on annual-average PM2.5
 5    concentrations.  Furthermore, reductions in both categories of risk, as we consider simulating just
 6    meeting current and alternative suites of standards, also depend primarily on changes in annual-
 7    average PM2.5 concentrations.
 8           The role of annual-average ambient PM2.5 concentrations in driving long-term exposure-
 9    related risk is intuitive given that this risk category is modeled using the annual-average air
10    quality metric.19  The fact that short-term exposure-related risk is also driven by changes in long-
11    term average PM2 5 concentrations is less intuitive, since changes in average daily PM2 5
12    concentrations are used to estimate changes in risk for this category.20 Analyses in previous PM
13    NAAQS risk assessments have shown that short-term exposure-related risks are not primarily
14    driven by the small number of days with PM2.5 concentrations in the upper tail of the air quality
15    distribution, but rather by the large number of days with PM2.5 concentrations at and around the
16    mean of the distribution. Consequently, consideration for changes in annual-average PM2 5
17    concentrations will explain to a large extent changes in short-term exposure-related risk.
18    Therefore, in interpreting patterns of long-term exposure-related risk, and the similar patterns we
19    observe in short-term exposure-related risk, we consider how simulating just meeting specific
20    suites of PM2.5 standards impacts the annual-average PM2.5 concentration for the study areas.
21           We have considered a series of questions to inform our understanding of the adequacy of
22    the current suite of fine particle standards based on the insights obtained from the quantitative
23    risk assessment. The remainder of the discussion is organized around these questions.
24    •   What is the nature and magnitude of the long-term exposure-related risks remaining
25        upon just meeting the current suite of PM2.s standards? What level of confidence is
26        associated with these risk estimates?
27           Of the fifteen study areas included in the risk assessment, thirteen are simulated to
28    experience risk reductions upon meeting the current suite of standards. Of these thirteen study
      19 As noted in section 3.2.1 of the second draft RA, estimates of long-term exposure-related mortality are actually
      based on an average annual PM25 level across monitors in a study area (i.e., the composite monitor annual-average).
      Therefore, in considering changes in long-term exposure-related mortality, it is most appropriate to compare
      composite monitor estimates generated for a study area under each suite of standards. The maximum monitor
      annual-average for a study area (i.e., the annual design value) determines the percent reduction in PM25 levels
      required to attain a particular standard. Both types of air quality estimates are provided in Tables F-49 and F-50 in
      Appendix F of the second draft RA and both are referenced in this discussion of core risk estimates, as appropriate.
      20 Estimates of short-term exposure-related mortality and morbidity are based on composite monitor daily PM2 5.
      concentrations. However, similar to the case with long-term exposure-related mortality, it is the maximum monitor
      98th percentile 24-hour concentration (the 24-hour design value) that will determine the degree of reduction required
      to meet a given 24-hour standard.
             Draft Do Not Quote or Cite            2-44                        March 2010

-------
 1    areas, three areas (Atlanta, Birmingham and Houston) have design values which resulted in the
 2    current annual standard controlling.  Therefore, in assessing the level of long-term exposure-
 3    related mortality risk remaining upon just meeting the current annual standard, we focus on risk
 4    estimates for these three study areas. Total incidence of PM2.5-related mortality ranges from
 5    131-165 (Birmingham) to 344-434 (Houston) (Table 2-2). The percent of total incidence of IHD
 6    mortality attributable to PM2.5 ranges from 10.7-13.6% (Houston) to 13.2-16.7% (Atlanta) (Table
 7    2-2). Total incidence estimates for all-cause and cardiopulmonary mortality related to PM2.5
 8    exposure for these study areas are larger than for IHD, while total PM2.5-attributable incidence
 9    estimates for lung-cancer are lower. However, the percent of total incidence attributable to PM2.5
10    exposure is larger for IHD than for the other mortality categories assessed.
11           The remaining ten study areas that would experience reductions in risk under the current
12    suite of standards  (relative to risk under recent conditions) have design values which result in the
13    current 24-hour standard controlling. These study areas include: Baltimore, Detroit, Fresno,  Los
14    Angeles, New York, Philadelphia, Pittsburgh, Salt Lake City, St. Louis, and Tacoma. Therefore,
15    we consider risk estimates for these study areas in assessing the degree of risk associated with
16    just meeting the current 24-hour standard.  Total incidence of PM2.s-related mortality ranges
17    from 15-19 (Salt Lake City) to  1,755-2,222 (New York City) (Table 2-2). The percent of total
18    incidence of IHD  mortality attributable to PM2.5 ranges from 2.9-3.7 (Salt Lake City) to 11.2-
19    14.2% (St. Louis) (Table 2-2). Total incidence estimates for all-cause and cardiopulmonary
20    mortality related to PM2.5 exposure for these study areas are larger than for IHD, while total
21    PM2.5-attributable incidence estimates for lung-cancer are lower. However, the percent of total
22    incidence attributable to PM2 5exposure is larger for IHD than for the other mortality categories
23    assessed.
24           Estimates of risk under the current suite of standards is significantly more variable for the
25    ten study areas where the 24-hour standard is controlling compared with the three study areas
26    where the current  annual standard is controlling. This notable difference in variability reflects the
27    fact that simulation of just meeting the current 24-hour standard (for those study areas where it is
28    controlling) produces varying impacts on annual-average PM2.5 concentrations. By contrast,
29    simulation of just  meeting  the annual standard results in similar annual-average PM2 5
30    concentrations for those study areas where  the annual standard is controlling. Because annual-
31    average PM2.5 concentrations are one of the primary determinants of the level of long-term
32    exposure-related mortality risk, variation in this statistic translates into greater variation in risk
33    remaining upon simulation of the current suite of standards for the ten study areas where the 24-
34    hour level is controlling.
35           A related point, which has bearing on the level of confidence associated with our risk
36    estimates, is the observation that several of the study areas where the 24-hour standard was
             Draft Do Not Quote or Cite           2-45                        March 2010

-------
1     Table 2-2. Estimated Incidence and Percent of Total Annual Incidence Associated with
2     Long-term PM2.s Exposure Based on Simulation of the Current Suite of Standards (for
3                    IHD mortality based on 2007 PM2.5 Concentrations)1'2
Risk Assessment
Location
Atlanta, GA
Baltimore, MD
Birmingham, AL
Dallas, TX
Detroit, Ml
Fresno, CA
Houston, TX
Los Angeles, CA
New York, NY
Philadelphia, PA
Phoenix, AZ
Pittsburgh, PA
Salt Lake City, UT
St. Louis, MO
Tacoma, WA
Design Values
(ug/m3)
Annual
16.2
15.6
18.7
12.8
17.2
17.4
15.8
19.6
15.9
15
12.6
19.8
11.6
16.5
10.2
24 -hour
35
37
44
26
43
63
31
55
42
38
32
60
55
39
43
Incidence of Ischemic Heart
Disease Mortality Associated
with Long-term Exposure to
PM253
Exposure
Period: 1979-
1983
220
(180-258)
297
(243 - 349)
131
(107-154)
195
(159-230)
377
(308 - 445)
77
(63 - 92)
344
(281 - 405)
860
(701 - 1018)
1755
(1435-2070)
261
(214-308)
317
(258 - 374)
256
(209 - 302)
15
(12-18)
446
(365 - 525)
38
(31 - 46)
Exposure
Period: 1999-
2000
277
(227 - 324)
	 374 	
(307 - 440)
165
(135-194)
247
(202-291)
478
(390 - 563)
98
(80- 116)
434
(355-511)
1094
(890-1296)
	 2222 	
(1814-2620)
330
(270 - 389)
402
(327 - 476)
324
(264 - 382)
19
(16-23)
563
(461 - 662)
49
(40 - 58)
Percent of Incidence of
Ischemic Heart Disease
Mortality Associated with Long-
term Exposure to PM2 /
Exposure
Period: 1979-
1983
10.4%
(8.5% -12.3%)
	 8.9% 	
(7.3% -10.5%)
8.4%
(6.9% - 9.9%)
9%
(7.3% -10.6%)
6.7%
(5.4% - 7.9%)
4.6%
(3.7% - 5.5%)
8.2%
(6.7% - 9.7%)
4%
(3.3% - 4.8%)
	 6.8% 	
(5.6% -8.1%)
7.9%
(6.4% - 9.3%)
5.8%
(4.7% - 6.8%)
6.9%
(5.6% -8.1%)
1 .3%
(1.1%- 1.6%)
8.6%
(7%- 10.1%)
2%
(1 .6% - 2.4%)
Exposure
Period: 1999-
2000
13.2%
(10.8% -15.5%)
113% 	
(9.2% -13.3%)
10.7%
(8.7% -12.6%)
1 1 .4%
(9.3% -13.4%)
8.5%
(6.9% -10%)
5.9%
(4.8% - 7%)
10.5%
(8.5%- 12.3%)
5.2%
(4.2% -6.1%)
	 87% 	
(7.1% -10.3%)
10%
(8.1% -11. 8%)
7.3%
(6% - 8.7%)
8.7%
(7.1% -10.3%)
1 .7%
(1 .4% - 2%)
10.9%
(8.9% - 12.8%)
2.5%
(2.1% -3%)
1The current primary PM2S standards include an annual standard set at 1 5 ug/m3 and a daily standard set at 35 ug/m3
2Nurnbers rounded to the nearest whole number. Numbers in parentheses are 95% confidence or credible intervals based on
statistical uncertainty surrounding the PM coefficient.
Estimates Based on Krewski et al. (2009), Using Ambient PM2js from1979-1983andfrom1999-2QOO respectively.
          Draft Do Not Quote or Cite
2-46
March 2010

-------
 1    controlling have annual-average PM2.5 concentrations that are well below the level of the current
 2    annual standard and in some instances, approaching the lowest measured level (LML) of 5.8
 3    Hg/m3 used as the lower bound for risk estimation. Although we are not able to quantify the
 4    degree of increased uncertainty, we do have reduced confidence in risk estimates involving
 5    annual-average PM2.5 concentrations that are substantially below the mean annual PM2.5
 6    concentrations reported by Krewski et al. (2009) (14-21 |ig/m3 - see Table 1 in the study) and
 7    approaching the LML of 5.8 |ig/m3 identified for that study.  This reflects the fact that, while
 8    available literature does not support a PM2.s-related threshold for health effects, we do have
 9    increased uncertainty in characterizing the nature of the C-R function as we move away from the
10    central mass of observations in the epidemiological study from which the C-R functions was
11    obtained and towards the LML. Study areas with relatively low annual-average PM2 5
12    concentrations less than the level of the current standard include: Tacoma (8.4 |ig/m3), Salt Lake
13    City (7.7 |ig/m3), and Fresno (9.9 |ig/m3) (note, that the current 24-hour standard is controlling
14    for all of these study areas - see US EPA, 2010a, Appendix F,  Table F-49). Conversely, we note
15    that we would have greater confidence in applying C-R functions in the risk assessment for those
16    study areas with annual-average PM2.5 levels under the current suite of standards near the range
17    of 14-21 |ig/m3.  Study  areas in this category include: Atlanta, Birmingham and Houston (where
18    the annual standard is controlling) and Baltimore, Detroit, Philadelphia, St. Louis, New York
19    (where the 24-hour standard is controlling).
20          Consideration of variability in the spatial pattern of PM2.5 reductions associated with
21    simulation of the current suite of standards suggests that this factor can impact risk estimates,
22    particularly in those instances where the 24-hour standard is controlling. For example,
23    consideration of a more localized pattern of reduction in PM2.5 concentrations (as  reflected in the
24    peak shaving rollback method), resulted in risk estimates for the 10 study areas where the 24-
25    hour standard is controlling that are from ~0 to 53% higher than those estimated assuming a
26    more regionalized pattern of PM2 5 reductions (as reflected in the proportional rollback method
27    used in the core analysis).
28          Additional sensitivity analyses considering sources of uncertainty impacting the  core risk
29    estimates focused on specification of the C-R function for long-term PM2 5 exposure-related
30    mortality.  This analysis suggested that most of the alternative model specifications supported by
31    available literature would produce risk estimates that were higher (by up to a factor of 2 to 3)
32    than the core risk estimates. These findings would apply both to estimates of PM2.5-attributable
33    IHD mortality incidence, as well as to estimates of the percent  of total IHD mortality incidence
34    attributable to PM2.5 exposure.
35          Taken together,  the sensitivity analyses completed for this risk assessment, including
36    those considering variability in rollback methods as well as uncertainty in the form of C-R
            Draft Do Not Quote or Cite           2-47                        March 2010

-------
 1    functions, suggest that the set of alternative risk model specifications that we identified generally
 2    produced risk estimates that are higher than the core risk estimates. Furthermore, our decision to
 3    model risk down to the LML (rather than to lower policy-relevant background (PRB)
 4    concentrations) for long-term PM2.5 exposure-related mortality, despite the lack of evidence for a
 5    threshold, results in lower estimates of risk that would have resulted from modeling risk down to
 6    PRB.  These considerations increase our overall confidence that we did not over-state risks with
 7    the core risk estimates. However, as noted above, there are different levels of confidence
 8    associated with risk estimates reflecting the annual-average PM2.5 concentrations associated
 9    with the risk estimates (with lower confidence associated with annual-average PM2.5
10    concentrations closer to the LML of 5.8 |ig/m3).
11    •  What is the nature and magnitude of the short-term  exposure-related risks remaining
12       upon just meeting the current suite of PM2.s standards? What level of confidence is
13       associated with these risk estimates?
14           As noted earlier, changes in annual-average PM2 5 levels drive reductions in both long-
15    term and short-term exposure-related risk. Consequently, patterns of risk reduction for long-term
16    exposure-related mortality  associated with the suite of annual standard levels generally hold for
17    short-term exposure-related mortality and morbidity (although absolute levels of risk will differ).
18    As was done above for long-term exposure-related mortality, we have divided the discussion of
19    risk for short-term exposure-related mortality and morbidity, between (a) the three study areas
20    where the current annual standard was driving and (b) the ten study areas where the 24-hour
21    standard was controlling. Again, this reflects the fact that these two groups of study areas had
22    different patterns of risk reduction under the current suite of standards, reflecting differences in
23    the way the two standard levels effected annual-average PM2.5 levels (i.e., fairly similar annual-
24    average PM2.5 levels for study areas where the annual standard controls and more variable
25    annual-average PM2.5 levels for study areas where the 24-hour standard controls).
26           Total incidence of short-term exposure-related CV mortality attributable for PM2.5 for the
27    three study  areas where the current annual standard controls (Atlanta, Birmingham and Houston)
28    ranges from 32 (Atlanta) to 46 (Houston) (Table 2-3). The percent of total CV mortality
29    incidence attributable to PM2 5 for this subset of study areas ranges from 0.8% (Atlanta) to 0.9%
30    (Houston) (Table 2-3). Total  incidence of CV hospital admissions (HA) attributable to PM2 5 for
31    this group of study areas ranges from 16 (Birmingham) to 56 (Houston) (Table 2-3), while
32    percent of total incidence attributable to PM2 5 ranges from 0.3% (Birmingham) to 0.4%
33    (Atlanta) (Table 2-3).
34           Total incidence of short-term exposure-related CV mortality attributable for PM2.5 for the
35    ten study areas where the 24-hour standard is controlling (Baltimore, Detroit, Fresno, Los
36    Angeles, New York, Philadelphia, Pittsburgh, Salt Lake City, St. Louis, and Tacoma), range
            Draft Do Not Quote or Cite            2-48                        March 2010

-------
 1    from 9 (Salt Lake City) to 106 (St. Louis) (Table 2-3). The percent of total incidence of CV
 2    mortality attributable to PM2.5 ranges from 0.7% (Tacoma) to 2.1% (New York and Philadelphia)
 3    (Table 2-3). Total incidence of CV hospital admissions (HA) attributable to PM2 5 for this group
 4    of study areas ranges from 9 (Salt Lake City) to 752 (Los Angeles) (Table 2-3), while percent of
 5    total incidence attributable to PM2.5 ranges from 0.4% (Salt Lake City) to 1.3% (St. Louis and
 6    Philadelphia) (Table 2-3).
 7           As observed with the long-term exposure-related mortality, estimates of short-term
 8    exposure-related mortality and morbidity risk under the current suite of standards are
 9    significantly more variable for the ten study areas where the 24-hour standard is controlling
10    compared with the three study areas where the current annual standard is controlling. Again, this
11    notable different in variability reflects the fact that simulation of just meeting the current 24-hour
12    standard (for those study areas where it is controlling) produces varying impacts on annual -
13    average PM2.5 concentrations and hence on risk. By contrast, those study areas where the annual
14    standard is  controlling have similar annual-average PM2.5 levels and consequently similar levels
15    of risk remaining upon simulation of the current suite of standards (recall that short-term
16    exposure-related risk is driven more by changes in annual-average PM2.5 then by  changes in peak
17    daily PM2.5 levels).
18           Given increased emphasis placed in this analysis on long-term exposure-related
19    mortality, the uncertainty analyses completed for this health endpoint category are more
20    comprehensive than those conducted for short-term exposure-related mortality and morbidity,
21    which to  some extent reflects limitations in study data available for addressing uncertainty in the
22    later category.
23    •  What roles do the current 24 hour and annual standards have in simulating the risks
24       remaining upon just meeting the current suite of standards?
25           As note above, of the thirteen urban study areas that do not meet the current suite of
26    standards based on 2005-2007 air quality data, ten areas have the 24-hour standard controlling,
27    while only 3 areas have the annual standard controlling. This pattern is generally characteristic of
28    the larger set of urban areas across the U.S. that do not meet the current suite of standards (US
29    EPA, 2010a, section 4.5).21
30           Estimated risks remaining upon just meeting the current suite of standards vary across
31    study areas, even when considering risks normalized for differences in population size and
32    baseline incidence rates.  This variability in estimated risks is a consequence of the variability in
33    the annual-average PM2.5 concentrations across study areas that result from simulating just
      21 Of the 256 urban areas in the U.S. (including a combination of CSA and CBSAs), 67 have ambient PM25 levels
      exceeding either the 24-hour or annual standard (or both) based on PM2 5 monitoring data from 2005-2007. Of these,
      51 have the 24-hour standard controlling, while 16 have the annual standard controlling.
             Draft Do Not Quote or Cite           2-49                         March 2010

-------
1       Table 2-3. Estimated Incidence and Percent of Total Annual Incidence Associated with
2      Short-Term PM2.s Exposure Based on Simulation of the Current Suite of Standards (C V
3               mortality and hospital admissions based on 2007 PM2.s concentrations)1'2
Risk Assessment
Location
Atlanta, GA
Baltimore, MD
Birmingham, AL
Dallas, TX
Detroit, Ml
Fresno, CA
Houston, TX
Los Angeles, CA
New York, NY
Philadelphia, PA
Phoenix, AZ
Pittsburgh, PA
Salt Lake City, UT
St. Louis, MO
Tacoma, WA
Design Values (ug/rn3)
Annual
16.2
15.6
18.7
12.8
17.2
17.4
15.8
19.6
15.9
15
12.6
19.8
11.6
16.5
10.2
24-hour
35
37
44
26
43
63
31
55
42
38
32
60
55
39
43
PM; s-Attriluilable Incidence
Cardiovascular
Mortality3
27
(-28-81)
54
(-4- 111)
-1
(-36 - 34)
29
(-19-76)
55
(-7-116)
12
(-9 - 33)
39
(-27-104)
-30
(-132-72)
461
^69 -651)
79
(20-136)
64
(-4-170)
42
	 (-BjJli 	
9
(-2 - 20)
92
(20-162)
11
(-6 - 27)
Cardiovascular
Hospital
Admissions4
35
(-23 - 92)
189
(139-239)
	 14 	
(-9-36)
28
(-18-73)
214
(157-271)
23
(0-46)
48
(-31 - 127)
	 258 	
	 S.:.51.1J 	
733
(538 - 927)
190
(140-240)
103
(1-215)
136
(100-172)
	 9 	
(0-18)
155
(114-196)
19
(-46 - 82)
Percent of Total Incidence
Attributable to PM2 5
Cardiovascular
Mortality3
0.7%
(-0.7% -2.1%)
1.4%
(-0.1% -2. 8%)
	 o% 	
(-1 .3% - 1 .2%)
0.8%
(-0.5% - 2.2%)
0.9%
(-0.1% -2%)
0.7%
(-0.5% - 2%)
0.8%
(-0.5% - 2%)
	 -a'2% 	
(-0.7% - 0.4%)
2%
(1.2% -2. 9%)
2%
(0.5% - 3.4%)
1 .3%
(-0.1% -2. 7%)
1%
(-0.2% - 2.3%)
	 678% 	
(-0.2% - 1 .7%)
	 'i76% 	
(0.4% - 2.9%)
0.7%
(-0.4% - 1 .8%)
Cardiovascular
Hospital
Admissions4
0.31 %
(-0.2% - 0.83%)
1.16%
(0.85% -1.47%)
	 6727% 	
(-0.1 8% -0.72%)
0.28%
(-0.1 8% -0.74%)
1.04%
(D.76%-1.32%)
0.48%
(0.01% -0.94%)
0.27%
(-0.1 7% -0.71%)
6746% 	
(D.01% - 0.91 %)
1.12%
(0.83% -1.42%)
1.18%
(D.87% - 1 .5%)
0.5%
(0.01% -0.99%)
1.08%
(0.8% -1.37%)
	 6736% 	
(D% - 0.7%)
	 i7i"% 	
(0.81 % - 1 .4%)
0.52%
(-1 .28% - 2.26%)
      The current primary
      2Percents rounded to the nearest tenth. Numbers in parentheses are 95% confidence or credible intervals based on statistical
      uncertainty surrounding the PM coefficient.

      3Based on location-specific single pollutant concentration-response function estimates from Zanobetti and Schwartz (2009) that have
      been "shrunken" towards the appropriate regional means. "Shrunken" coefficient estimates and their standard errors were sen

      Incidence estimates were calculated using the appropriate regional concentration-response function estimates reported in Table 2 of
      Bell et al. (2008). Location-specific C-R function estimates were not available from this study.
             Draft Do Not Quote or Cite
2-50
March 2010

-------
 1    meeting the current standards.22 Variability in annual-average PM2.5 concentrations is
 2    substantially greater in those study areas in which the 24-hour standard is controlling. In such
 3    areas, the variability across study areas in estimated risks is largest when regional patterns of
 4    reductions in PM2.5 concentrations are simulated (using proportional rollback, as was done in the
 5    core analyses), with less variability when more localized patterns of PM2.5 reductions are
 6    simulated (using peak shaving rollback, as was done in a sensitivity analysis).
 7           In simulating just meeting the current suite of standards for the urban study areas, the
 8    resulting annual-average PM2.5 concentrations  range from about 15 |ig/m3 (for the study areas in
 9    which the annual standard was controlling) down to as low as about 8 |ig/m3 (for the study areas
10    in which the 24-hour standard was controlling). We note that, for long-term exposure-related
11    mortality, estimates of risk based on annual-average PM2 5 concentrations approaching the LML
12    of 5.8 |ig/m3 have greater uncertainty than estimates based on annual-average concentrations
13    closer to the mean of the epidemiological study from which the C-R functions are obtained (14 to
14    21 |ig/m3).23 Therefore, we generally have less confidence in risk estimates generated for study
15    areas where the 24-hour standard is controlling since these tend to have lower annual-average
16    PM2.5 levels (which can approach the LML) compared with study areas where the annual-
17    average is controlling.
18           These observations when considered together suggest that, when we simulate PM2.5
19    concentrations meeting the current suite of standards, we find that in study areas where the
20    current 24-hour standard is controlling, the degree of public health protection afforded by the
21    current suite of standards is much more variable than in study areas where the annual standard is
22    controlling. Furthermore, we generally have less confidence in the risk estimates generated for
23    urban areas where the current 24-hour standard is controlling due simulated lower annual-
24    average PM2.5 concentrations that are in some cases much lower, approaching the LML.
25    •  How representative are the risk estimates generated for the urban study areas from a
26       national perspective?
27           As part of the risk assessment, we completed several additional analyses intended to
28    place  the core risk estimates in the broader national-context by considering the degree to which
29    the 15 urban study areas are representative of larger urban areas within the U.S., particularly
30    areas likely to experience elevated risk related  to PM exposure. Below we provide brief
31    descriptions of each analysis, along with observations results from each analysis regarding the
      22 As noted earlier, changes in both short-term and long-term exposure-related risk reflect primarily changes in long-
      term average (annual) PM2 5 levels.
      23 This observation also holds for estimates of risk associated with short-term PM2 5 although we note that the LML
      for these studies is below PRB, which allows us to make estimates of risk for lower PM2 5 levels with greater
      confidence relative to modeling long-term exposure-related mortality with its higher LML.
            Draft Do Not Quote or Cite            2-51                          March 2010

-------
 1    representativeness of the urban study areas:

 2    •   The representativeness analysis (US EPA, 2010a, section 4.4) compared attributes of the 15
 3        urban study eras (assessed at the county-level) against national distributions for the same
 4        attributes. The analysis suggests that the 15 urban study areas represent areas in the U.S. that
 5        are among the most densely populated, have relatively higher levels of annual and 24-hour
 6        98th percentile PM2.5 concentrations, and capture well the range of effect estimates
 7        represented by the Zanobetti and Schwartz (2009) study.  Together, these factors suggest that
 8        the urban study areas should capture well the overall distribution of risk for the nation, with
 9        the potential for better characterization of the high end of that distribution.24

10    •   Consideration of the mix of design values across the 15 urban study areas as contrasted with
11        design values for the broader set of urban study  areas in the U.S. suggests that the 15 urban
12        study areas do a good job of capturing the key groupings of urban areas in the U.S. likely to
13        experience elevated risk due to PM (i.e., we have coverage for each of the zones containing
14        urban study areas likely to experience risk reductions under the suites of alternative standard
15        levels considered - see US EPA,2010a, section 4.5.1). Furthermore, this analysis suggested
16        that we have also included study areas likely to experience relatively greater degrees of
17        PM2.5-related risk, considering the pattern of design values across urban areas in the U.S.

18    •   Consideration of where the 15 urban study areas fell along the distribution of U.S. counties
19        included in the national-scale mortality analysis further suggests that we have captured
20        counties likely to experience elevated PM25-related risk.  As part of the national-scale
21        mortality analysis (see US EPA, 2010a, chapter 5), we created a cumulative distribution of
22        the percentage of mortality attributable to PM2.s based on the county-level estimates for the
23        U.S.25 We then identified where along this cumulative distribution the 31  counties
24        comprising our 15 urban study areas fell.  This analysis suggests that our urban study areas
25        capture the upper end of the tail with regard to PM2.5-attributable risk, with 23 of these
26        counties falling within the upper 5th percentile of the distribution.  These findings support the
27        assertion based on the other analyses described above that the urban study areas are likely to
28        capture risk at urban areas experiencing relatively elevated levels of PM2 5_attributable
29        mortality.
30           Our overall assessment of the representativeness of the 15 urban study areas in the
31    national context, based on the three analyses summarized above, is that our study areas do a good
32    job of representing urban areas in the U.S. experiencing elevated levels of risk related to ambient
33    PM2 5 exposure. The results of the national-scale mortality analysis also  suggest that, while our
34    15 urban study areas do provide coverage for urban areas in the U.S. experiencing elevated
35    levels of PM2.5-related risk, there are many additional areas (counties) not modeled in the risk
      24 This analysis also showed that the urban study areas do not capture areas with the highest baseline morality risks
      or the oldest populations (both of which can result in higher PM2 5-related mortality estimates). However, some of
      the areas with the highest values for these attributes have relatively lower PM2 5 levels (e.g., urban areas in Florida)
      and consequently failure to include these areas in the set of urban study areas is unlikely to bias the risk estimates in
      terms of excluding high PM2 5-risk locations.
      25 Note that by using this risk metric, we avoid influence by difference in overall population size (as would be the
      case with raw incidence) and focus on a unitized estimate of PM2 5-related mortality which reflects differences in (a)
      baseline mortality incidence, and (b) the annual average PM2 5 concentrations for each county.

             Draft Do Not Quote or Cite           2-52                         March 2010

-------
 1    assessment that experience elevated PM2.5-related risk.  In other words, it should not be
 2    construed that significant PM2.5-related risk is limited only to the urban study areas included in
 3    the risk assessment.
 4    •  To what extent are the risks remaining upon simulation of the current suite of
 5       standards important from a public health perspective?
 6          With respect to considering the results of the quantitative risk assessment to inform our
 7    understanding of risks associated with long-term PM2 5 exposures, we reach the preliminary
 8    conclusion that the long-term exposure-related mortality risks remaining upon simulation of just
 9    meeting the current suite of PM2.5 standards can reasonably be judged to be important from a
10    public health perspective. This preliminary conclusion is based, in part, on both the nature of the
11    risks (total and cause-specific mortality) as well as the magnitude of these risks (the total
12    incidence and percentage of incidence associated with PM2.5 exposure as presented in Table 2-2).
13    In addition, based on consideration of both our qualitative and quantitative assessments of
14    uncertainty, as well as additional factors considered in the risk model (e.g., modeling risk down
15    to LML rather than PRB), we are reasonably confident that we have not overstated the
16    magnitude of risk associated with simulating the current suite of standards. However, we do note
17    that we have greater overall confidence in estimates of risk associated with long-term PM2.5
18    exposure in the subset of urban study areas for which the current annual standard is controlling,
19    since these locations tend to have higher annual-average PM2.5 concentrations. In contrast, the
20    urban study areas where the current 24-hour standard is controlling tend to have substantially
21    lower annual-average PM2.5 concentrations, which reduces our overall confidence in these
22    estimates of risks related to long-term exposures.
23          With regard to considering the results of the quantitative risk assessment to inform our
24    understanding of risks associated with short-term PM2.5 exposures, we reach the preliminary
25    conclusion that short-term exposure-related mortality and morbidity risks upon simulation of the
26    current suite of PM2.5 standards can reasonably be judged to be important from a public health
27    perspective.  Similar to the discussion of risks associated with long-term PM2.5 exposure, this
28    preliminary conclusion is based, in part, on both the nature of the risks (total and cause-specific
29    mortality as well as cause-specific hospital admissions) as well as the magnitude of those risks
30    (total incidence and percentage of incidence associated with PM2.s exposure as presented in
31    Table 2-3). We note however, that in the case of mortality associated with short-term PM2 5
32    exposure, the magnitude of these risks is  substantially lower than that associated with long-term
33    exposure to PM2 5. With regard to our overall confidence in the estimates of risks related to
34    short-term PM2 5 exposures, while we believe that the core simulation of risk is based on C-R
35    functions that are well-supported in the literature,  we acknowledge that we have not completed
36    as comprehensive an assessment of uncertainty and variability as was done for estimates of
            Draft Do Not Quote or Cite          2-53                        March 2010

-------
 1    mortality risks related to long-term PM2.5 exposure. Therefore, our overall confidence that we
 2    have not over-stated risk associated with short-term exposure-related mortality and morbidity is
 3    not as high as for long-term exposure-related mortality, although we have no reason to believe
 4    that we have over-stated risk.

 5    2.2.3  Preliminary Staff Conclusions on Adequacy of Current Standards
 6          Collectively, taking into consideration the responses to specific questions focusing on
 7    different ways to address the adequacy of the current suite of PM2.5 standards, we revisit the
 8    overarching policy question: does the currently available scientific evidence and risk-based
 9    information support or call into question the adequacy of the protection afforded by the current
10    suite of fine particle standards?
11          With respect to evidence-based considerations, the currently available evidence provides
12    stronger evidence beyond what was available in the last review, that associations between short-
13    and long-term PM2.5 exposures and a broad range of adverse health effects exist. The newly
14    available information strengthens the associations between PM2 5 and mortality and
15    cardiovascular and respiratory morbidity effects observed in the  last review and expands our
16    understanding of a broader range of health outcomes as well as our understanding of effects in
17    susceptible populations.  The newly available evidence provides support for associations that
18    extend to lower concentrations than what had been observed in the last review, including at
19    ambient concentrations below the levels of the current standards.
20          In relation to risk-based considerations for informing our understanding of the adequacy
21    of the current fine particle standards, we focus on the estimates of PM2.5-related mortality and
22    morbidity effects likely to remain upon meeting the current standards in a number of example
23    urban areas.  In considering the core risk estimates together with our understanding of the
24    uncertainties in these estimates based upon extensive sensitivity  analyses, we reach the
25    preliminary conclusion that the risks estimated to be associated with just meeting the current
26    standards can reasonably be judged to be important from a public health perspective and these
27    estimated risks provide support for consideration of standards that would provide increased
28    protection beyond that afforded by the  current PM2 5 standards.
29          We recognize that important uncertainties and research questions  remain when
30    considering both evidence- and risk-based approaches.  Nonetheless, we note that much progress
31    has been made in reducing some key uncertainties since the last review, including important
32    progress in advancing our understanding of potential mechanisms by which ambient PM2.5 is
33    causally linked with mortality, cardiovascular, and respiratory effects observed in epidemiologic
34    studies.  Additional information continues to emerge for a broader range of health effects
35    including reproductive and development effects and more information is available to understand

            Draft Do Not Quote or Cite           2-54                        March 2010

-------
 1    susceptible populations including children, older adults, and individuals with pre-existing
 2    cardiovascular and respiratory disease.  As was true in the last review, we recognize that as the
 3    body of available evidence has expanded, it has added greatly both to our knowledge of health
 4    effects associated with fine particle exposures, as well as to the complexity inherent in
 5    interpreting the evidence in a policy-relevant context as a basis for setting appropriate standards.
 6    In evaluating both evidence-based and risk-based considerations, along with associated
 7    limitations and uncertainties, we reach the preliminary conclusion that the available information
 8    clearly calls into question the adequacy of the current suite of PM2.5 standards and provides
 9    strong support for giving consideration to revising the current standards to provide increased
10    public health protection.

11    2.3    CONSIDERATION OF ALTERNATIVE STANDARDS
12          Having reached the conclusion that the currently available scientific evidence calls into
13    question the adequacy of the current suite of PM2.5 standards, staff considers a second
14    overarching question:
15
16
17
  What alternative suite(s) of fine particle standards is (are) supported by the currently
available scientific evidence and risk-based information, as reflected in the ISA and second
                                       draft RA?
18           To inform the answer to this overarching question, we have posed a series of more
19    specific questions to aid in considering how the current suite of primary PM2.5 standards might be
20    revised to provide requisite public health protection.  Specifically, we consider how the currently
21    available scientific evidence informs decisions regarding the basic elements of the NAAQS:
22    indicator (section 2.3.1), averaging time (section 2.3.2), form (section 2.3.3), and level (sections
23    2.3.4 and 2.3.5). These elements will be considered collectively in evaluating the health
24    protection afforded by alternative suites of standards under consideration.  In considering the
25    currently available scientific and technical information, we consider both the information
26    available in the last review and information that is newly available since the last review as
27    assessed and presented in the ISA and second draft RA prepared for this review (US EPA,
28    2009a;USEPA, 2010a).

29    2.3.1   Indicator
30           In 1997, EPA decided that particles from the fine and coarse fractions of thoracic
31    particles (PMio) should be defined as separate pollutants. At that time,  the Agency established
32    PM2.5 as the indicator for fine particle, while retaining PMio as the indicator for thoracic coarse
33    particles. In determining the indicator for fine particles, the Agency first considered whether the
34    indicator should be based on the mass of a size-differentiated sample of fine particles or one or

             Draft Do Not Quote or Cite           2-55                         March 2010

-------
 1    more components within the mix of fine particles. In that review, EPA first concluded it was
 2    more appropriate to control fine particles as a group, as opposed to singling out any particular
 3    component or class of fine particles.  Second, in establishing a size-based indicator, a size cut
 4    was selected that would appropriately distinguish fine particles from particles in the coarse
 5    mode. In focusing on a size cut within the size range of 1 to 3 um (i.e., the intermodal range
 6    between fine and coarse mode particles), EPA recognized that the choice of any specific
 7    sampling size cut within this range was largely a policy judgment.  In making this judgment, the
 8    Agency noted that the available epidemiologic studies of fine particles were based largely on
 9    PM2.5 and also considered monitoring technology that was generally available.  In the
10    Administrator's final decision, the selection of a 2.5  um size cut reflected the regulatory
11    importance that was placed on defining an indicator that would more completely capture fine
12    particles under all conditions likely to be encountered across the U.S., especially when fine
13    particle concentrations are likely to be high, while recognizing that some small coarse particles
14    would also be captured by current methods to monitor PM2.5 (62 FR 38667 to 38668, July, 18,
15    1997).
16          In reaching the decision to retain PM2.5 as the indicator for fine particles in the last
17    review, the same considerations continued to apply for selection of an appropriate  indicator for
18    fine particles (71 FR 61162 to 61164, October, 17, 2006). The available epidemiologic studies
19    linking mortality and/or morbidity effects with short- and long-term exposures to fine particles
20    continued to be largely indexed by PM2.5. While emerging evidence focused on various PM2.5
21    constituents and provided some evidence related to various components within the mix of fine
22    particles (e.g., sulfates (SO2"4), nitrates, elemental carbon (EC), organic compounds, and metals)
23    as well as associations between mortality and particles from different sources of fine particles,
24    this evidence was deemed too limited to support a distinct standard for a specific PM2 5
25    component or fine particle source.  More specifically, the Agency concluded, there was "not
26    sufficient evidence that would lead toward the selection of one or more PM components as being
27    primarily responsible for effects associated with fine particles, nor is there sufficient evidence to
28    suggest that any component should be eliminated from the indicator for fine particles" (71 FR
29    61163/1).
30          In this review, in considering  alternative fine particle standards, we first address the issue
31    of indicator by asking the following question:
32    •   Does the currently available evidence provide support for the continued use of PM2.s as
33       the mass-based indicator for fine particles?
34          As noted above, the selection  of PM2.5 to characterize respirable particles in 1997 was
35    driven mainly by considerations related to measurement techniques available at the time rather
36    than dosimetry. In this review, the ISA notes, "[Currently, cut points other than 2.5 um are
            Draft Do Not Quote or Cite           2-56                        March 2010

-------
 1    attainable and frequently put into use... [m]ost commonly, however, PM2.5 is used as an indicator
 2    of respirable particles" (US EPA, 2009a, p. 3-3). Recently, there has been increasing interest in
 3    examining the relationship between the particle number concentration by size and health effects.
 4    However, several instruments are needed to provide size distribution measurements (number and
 5    size) over the several orders of magnitude of particle diameters of interest (US EPA, 2009a,
 6    section 3.4.1.5, p. 3-29). These techniques, while widely used in aerosol research, have not yet
 7    been widely used in health effects studies.
 8           As discussed in section 2.2.1, currently available scientific information providing
 9    evidence of associations between fine particles and a broad range of health outcomes has been
10    substantially strengthened in this review.  Epidemiological evidence continues to provide the
11    strongest support for standards to protect public health from long- and short- term fine particle
12    exposures.  Measurements from community-based ambient monitors have generally been used
13    for time-series and longitudinal epidemiological studies, and may be used for panel studies as
14    well (US EPA, 2009a, p. 3-176). New epidemiological studies available in this review have
15    considered a much larger set of air quality data than was available in the last review based
16    primarily on ambient measurement data from the existing PM2.5 monitoring networks.26 This
17    includes consideration of air quality measurements reported in more recent years. These data
18    have improved our understanding of the spatiotemporal distribution of PM2.5 (US EPA, 2009a,
19    sections 3.5 and 3.9).
20           As presented in the ISA, epidemiological studies linking cardiovascular and respiratory
21    effects as well as  mortality with short- and long-term fine particle exposures continue to be
22    largely indexed by PM2 5 (US EPA, 2009a, chapters 6 and 7).  Thus, we reach the preliminary
23    conclusion that it is reasonable to retain PM2.5 as an indicator for fine particles.  We then pose
24    additional questions to aid in considering whether the currently available scientific evidence
25    provides support for supplementing the current PM2.5 mass-based indicator by considering
26    whether there is sufficient evidence to support a standard with a different size fraction or whether
27    there is sufficient evidence to establish  distinct standards focused on regulating specific PM2 5
28    components or sources of fine particles.
29    •  To what extent does the currently available information provide support for
30       considering a separate indicator for ultrafine particles as a subfraction of fine particles?
      26 The network of PM25 Federal Reference Methods (FRM) monitors has been operational since 1999. This network
      currently includes over 900 monitoring stations around the U.S. FRM-like air quality data are available from
      continuous monitors required in metropolitan statistical areas (MSAs) that have an FRM/Federal Equivalent (FEM)
      monitor.  The network of continuous PM2 5 monitors has grown to over 700 locations throughout the U.S. While
      PM2 5 continuous monitors primarily support forecasting and reporting for the Air Quality Index (AQI), they are also
      used in interpreting the diurnal characterization of PM2 5
             Draft Do Not Quote or Cite          2-57                         March 2010

-------
 1           In considering whether or not the currently available evidence provides support for an
 2    additional size-based indicator by focusing on ultrafme particles (UFPs)27, we recognize that a
 3    number of studies have focused on UFPs as a subset of PM2 5.  In the last review, limited
 4    evidence was available suggesting that the ultrafme subset of fine particles (generally including
 5    particles with a nominal aerodynamic diameter less than 0.1 jim) was associated with adverse
 6    health effects (US EPA, 2004, pp. 8-66 to 8-68). In this review, the ISA notes that there are
 7    many reasons for looking more closely at effects associated with  this size  fraction including, the
 8    particle number and large surface area of UFPs per unit of mass.  Particle  number is most highly
 9    concentrated in the UFP fraction with volume (or mass) most concentrated in the larger size
10    fractions (US EPA, 2009a, p. 3-2) Furthermore, per unit mass, UFPs may have more opportunity
11    to interact with cell surfaces due to their greater surface area and  their greater particle number
12    compared with larger particles (US EPA, 2009a, p. 5-3). Greater  surface area increases the
13    potential for soluble components to adbsorb to UFPs and be transported into the body (US EPA,
14    2009a, p. 6-83).  Many studies suggest that the surface of particles or substances released from
15    the surface (e.g., transition metals, organics) interact with biological substrates, and that surface-
16    associated free radicals or free radical-generating systems may be responsible for toxicity,
17    resulting in greater toxicity of UFPs per particle surface area than larger particles. In addition,
18    evidence available in this review suggests that the ability of particles to enhance allergic
19    sensitization is associated more strongly with particle number and surface area than particle mass
20    (US EPA, 2009a, p. 6-127). Evidence is also available in this review suggesting that smaller
21    particles may have a greater potential to cross cell membranes and epithelial barriers. The ISA
22    notes that, for UFPs, "enhanced translocation to interstitial compartments  or to the circultaon
23    may be important sequelae" (US EPA, 2009a, p. 5-6).  More information on possible modes of
24    action for effects associated with UFPs exposures is discussed in sections  5.1 and 5.4 of the ISA.
25           With respect to ambient concentrations of UFPs, at present, there is no national network
26    of UFP samplers, thus, only episodic and/or site-specific data sets exist (US EPA, 2009a, p. 2-
27    2).28  Therefore, a national characterization of concentrations, temporal and spatial patterns, and
28    trends is not possible, and the availability of ambient UFPs data to support health studies are
29    extremely limited.  In general, UFP particle number concentrations are highly  dependent on
30    monitor location and therefore, more subject to exposure error than accumulation mode particles
31    (US EPA, 2009a, p. 2-22). The UFP number concentrations fall off sharply downwind from
32    sources, as UFPs may grow into the accumulation mode by coagulation or condensation (US
      27 Ultrafine particles (UFPs) are emitted directly to the atmosphere or are formed by nucleation of gaseous
      constituents in the atmosphere (US EPA, 2009a, p. 3-3).
      28 The ISA contains a review of the current scientific information related to measurements of UFPs (US EPA,
      2009a, sections 3.5.1 and 3.5.2).
            Draft Do Not Quote or Cite          2-58                        March 2010

-------
 1    EPA, 2009a, p. 3-89). Limited studies of UFP ambient measurements suggest these particles
 2    exhibit a high degree of spatial and temporal heterogeneity driven primarily by differences in
 3    nearby source characteristics (US EPA, 2009a, p. 3-84). Internal combustion engines and
 4    therefore, on-road roadways are a notable source of UFPs, so concentrations of UFPs near
 5    roadways can generally be expected to be elevated (US EPA, 2009a, p. 2-3).  Concentrations of
 6    UFPs have been reported to drop off much more quickly with distance from roadways than
 7    larger particle sizes (US EPA, 2009a, p. 3-84).
 8          Additional but still limited health evidence  available in this review, primarily from
 9    controlled human exposure and toxicological studies, provide evidence for UFP-induced
10    cardiovascular and respiratory effects. This evidence is largely related to studies focused on
11    exposure to diesel exhaust (DE), "[a]s a result, it is unclear if the effects observed are due to
12    UFP, larger particles (i.e., PM2.5), or the gaseous components of DE" (US EPA, 2009a, p. 2-22).
13    The ISA notes uncertainty associated with the controlled human exposure studies as concentrated
14    ambient particle (CAP) systems have been shown to modify the composition of UFPs (US EPA,
15    2009a, p.  2-22, see also section 1.5.3). In addition, relatively few epidemiologic studies have
16    examined the potential cardiovascular and respiratory effects associated with  short-term
17    exposures to UFPs. These studies have reported inconsistent and mixed results (US EPA, 2009a,
18    section 2.3.5).
19          In considering the body of scientific evidence available in this review, the ISA concludes
20    that the currently available evidence is suggestive of a causal association between short-term
21    exposures to UFPs and cardiovascular and respiratory effects. Furthermore, the ISA concludes
22    that evidence is inadequate to infer a causal association between short-term exposure to UFPs
23    and mortality as well as long-term exposure to UFPs and all outcomes evaluated (US EPA,
24    2009a, sections 2.3.5, 6.2.12.3, 6.3.10.3, 6.5.3.3, 7.2.11.3, 7.3.9, 7.4.3.3, 7.5.4.3, and 7.6.5.3;
25    Table 2-6).  Thus, while new evidence expands our understanding of effects associated with
26    UFPs, we reach the preliminary conclusion that this information is still too limited to support a
27    distinct PM standard for UFPs.
28    •   To what extent does  the currently available information provide support for
29       considering a separate indicator for a specific PMi.5 component or source category of
30       fine particles?  Conversely, to what extent does the currently available information
31       provide support for  eliminating any component or source category from the mix of fine
32       particles included in the PMi.s indicator?
33          In the last review, EPA recognized the availability of a limited number of epidemiologic
34    studies that explored associations with various components within the mix of fine particles and
35    adverse effects (US EPA, 2004, section 9.2.2.1.1, Table 9-3) as well as several studies that used
36    PM2.5 speciation data to evaluate the association between mortality and particles from different

            Draft Do Not Quote or Cite          2-59                        March 2010

-------
 1    categories of fine particle sources (US EPA, 2004, section 8.2.2.5). As discussed in the last
 2    review, different patterns of associations of various PM2.5 components (e.g., sulfates, nitrates,
 3    metals, organic compounds, elemental carbon (EC)) or source categories of fine particles with
 4    total or cardiovascular mortality were observed in various  short-term exposure studies (US EPA,
 5    2004, section 8.2.2.5. Tables 8-3,  8-4, 9-3). However, as noted in the last review, "many PM
 6    components are correlated with each other and also with PM mass, making it difficult to
 7    distinguish effects of the various components. Also, different PM components or characteristics
 8    would be expected to be more closely linked with different health outcomes" (US EPA, 2004, p.
 9    9-30). In addition, a limited number of studies evaluated in the last review used PM2.5 speciation
10    data to assess the effects of air pollutant combinations or mixtures using factor analysis or source
11    apportionment methods to link effects with different PM2 5 source types (Schwartz, 2003; Mar et
12    al., 2003; and Tsai et al., 2000). These studies reported that fine particles from combustion
13    sources, including motor vehicle emissions, coal combustion, oil burning and vegetative burning,
14    were associated with increased mortality. No significant increase in mortality was reported with
15    a source factor representing crustal material in fine particles (US EPA, 2004, section 8.2.2.5.3).
16    The EPA concluded that these studies indicated that exposure to fine particles from combustion
17    sources, but not crustal material, was associated with mortality (US EPA,  2004, p. 8-85; US
18    EPA, 2005, p. 3-16).
19          Overall,  EPA concluded in the last review that the  available evidence suggested "that
20    many different chemical components of fine particles and  a variety of different types of source
21    categories are all associated with, and probably contribute  to, mortality, either independently or
22    in combinations" (US EPA, 2004,p. 9-31). Conversely, scientific evidence available in the last
23    review provided no basis to conclude that any individual fine particle component  could not be
24    associated with adverse health effects (US EPA, 2005, p. 5-17).  This evidence provided the
25    basis for EPA to reach the final  decision that "there was not sufficient evidence that would lead
26    EPA to select one or more PM2 5 component as being primarily responsible for effects associated
27    with fine particles, nor was there sufficient evidence to suggest that any component should be
28    eliminated from the indicator for fine particles" (71 FR 61163/1,  October  17, 2006).
29          In addressing the issue of particle composition in this review, the ISA concludes that,
30    "[f]rom a mechanistic perspective, it is highly plausible that the chemical  composition of PM
31    would be a better predictor of health effects than particle size"  (US EPA, 2009a, p. 6-202).
32    Heterogeneity of ambient concentrations of PM2 5 constituents (e.g., elemental carbon (EC),
33    organic carbon (OC), sulfate (SO42"), and nitrate)  observed in different geographical regions as
34    well as regional heterogeneity in PM2.5-related health effects reported in a number of
35    epidemiologic studies are consistent with this  hypothesis (US EPA, 2009a, section 6.6).

            Draft Do Not Quote or Cite           2-60                        March 2010

-------
 1           With respect to the availability of ambient measurement data for fine particle
 2    components, we recognize that, in this review, there are more extensive ambient PM2.5 speciation
 3    measurement data available through the Chemical Speciation Network (CSN).29 Data from the
 4    CSN monitoring network provide further evidence of spatial and seasonal variation in both PM2.5
 5    mass and composition among cities/regions (US EPA, 2009a, pp. 3-50 to 3-60; Figures 3-12 to
 6    3-18; Figure 3-47).  Some of this variation may be related to "regional differences in
 7    meteorology, sources, and topography" (US EPA, 2009a, p. 2-3).  While the network of
 8    approximately 200 CSN monitoring sites provides valuable data for development and tracking of
 9    control strategies, its use for providing PM2.5 speciation data to support epidemiological studies
10    is somewhat limited.  The CSN sites provide measurement data on a one-in-three or one-in-six
11    day schedule and do not capture data every day. Health researchers have expressed a strong
12    interest in having access to PM2.5 speciation measurements collected more frequently.30
13           The currently available epidemiologic, lexicological, and controlled human exposure
14    studies have evaluated the health effects associated  with ambient PM2.5 constituents and
15    categories of fine particle sources, using a variety of quantitative methods applied to a broad set
16    of PM2 5 constituents,  rather than selecting a few constituents a priori (US EPA, 2009a, p. 2-26).
17    Epidemiological studies have used measured ambient PM2.5 speciation data, including
18    monitoring  data from  the CSN, while all of the controlled human exposure and most of the
19    toxicological studies have used CAPs, and analyzed the constituents therein (US EPA, 2009a, p.
20    6-203).31
      29The CSN consists of 54 Speciation Trends Network (STN) sites as well as about 150 SLAMS supplemental sites
      across the country measuring over 40 chemical species. A limited number of CSN monitors began collecting
      ambient data in 2000 with the majority of sites collecting data starting in 2001. These sites collect aerosol samples
      over 24 hours on filters that are analyzed for PM2 5 mass, trace elements, major ions (e.g., sulfates, nitrates,
      ammonium), and organic and elemental carbon.
      30 As outlined in section 6.6.2.1 lof the ISA, some investigators have circumvented the issue of less than daily
      speciation data by using the PM2 5 chemical species data in a second stage regression to explain the heterogeneity in
      PM10 or PM2 5 mortality risk estimates across cities and assuming that the relative contributions of PM25 have
      remained the same over time (US EPA, 2009a, p. 6-206). In April 2008, EPA co-sponsored a workshop to discuss
      modifications to the current ambient air quality monitoring networks that would advance our understanding of the
      impacts of PM exposures on public health/welfare in the most meaningful way, including improving our
      understanding of components of fine particles. This workshop was a major step in a series of interactions to foster
      improved long-term communication between external stakeholders, including air quality monitoring experts and
      health researchers. A summary of the workshop recommendations, including recommendations for daily PM25
      speciation measurements in large urban areas, is available at www.epa.gov/ORD/npd/pdfs/FINAL-April-2008-AQ-
      Health-Research-Workshop-Summary-Dec-2008.pdf.  Follow-up for a number of the workshop recommendations is
      on-going.
      31 Most studies considered between 7 and 20 ambient PM2 5 constituents, with EC, OC, SO42", nitrate, and metals
      most commonly measured. Many of the studies reduced the  number of ambient PM2 5 constituents by grouping them
      with various factorization or source apportionment techniques to examine the relationship between the grouped
      PM2 5 constituents and various health effects. However, not all studies labeled the constituent groupings according to
      their presumed source and a small number of controlled human exposure and toxicological studies did not apply any
      kind of grouping to the ambient PM2 5 speciation data.  In addition, there were differences in the type and grouping
             Draft Do Not Quote or Cite            2-61                          March 2010

-------
 1           With respect to epidemiological studies evaluating short-term exposures to fine particle
 2    constituents, several new multi-city studies are now available.  These studies continue to show an
 3    association between mortality and cardiovascular and/or respiratory morbidity effects and short-
 4    term exposures to various PM2.5 components including nickel (Ni), vanadium (V), elemental
 5    carbon (EC), organic carbon (OC), and sulfates (US EPA, 2009a, sections 6.5.2.5 and  6.6).
 6    Lippmann et al. (2006); Dominici et al. (2007) evaluated the heterogeneity in the PMi0-mortality
 7    association as evaluated in the NMMAPS data by analyzing the PM2.5 speciation data.  Nickel
 8    (Ni) and Vanadium (V) were identified as significant predictors of variation in PMio-related
 9    mortality across cities, with Ni levels in New York City being reported as particularly  high (US
10    EPA, 2009a, section 6.5.2.5; Figure 6-31).32 Bell et al. (2009) and Peng et al. (2009) conducted
11    similar analyses focusing on the variation in PM25-related cardiovascular and respiratory
12    hospital admissions in older adults.  Peng et al. (2009) focused on the components that make up
13    the majority of PM2 5 mass and  found that in multi-pollutant models only EC and OC were
14    significantly associated with risk of hospitalization for cardiovascular disease. Bell et al. (2009)
15    used data from twenty PM2.5 components and found that EC, Ni, and V were most positively and
16    significantly associated with the risk of PM2.s-related hospitalizations suggesting that the
17    observed associations between PM2.5 and hospitalizations may be primarily due to particles from
18    oil combustion and traffic (US EPA, 2009a, section 6.2.10.1). In a study of 25 U.S. cities,
19    Franklin et al. (2008) focused on a time-series regression of mortality related to PM2 5 mass by
20    season and also examined effect modification due to various PM2.5 species. They concluded that
21    Al, As, Ni, Si and SC>42" were significant effect modifiers of PM2 5 mortality risk estimates, and
22    "simultaneously including Al, Ni, and SC>42" together or Al, Ni, and As together further increased
23    explanatory power. Of the species examined, Al and Ni explained the most residual
24    heterogeneity" (US EPA, 2009a, p. 6-194; Table 6-17).33  Furthermore, Ostro et al (2006)
25    examined associations between PM2.5 components and mortality in six California counties and
26    found an association between mortality, especially cardiovascular-related mortality and several
27    PM2.s components including EC, OC, nitrate, Fe, K, and Ti at various lags (US EPA, 2009a, p. 6-
28    195).
      of PM25 constituent data used in the various studies and analyses conducted resulting in important limitations in
      interpreting these studies together (US EPA, 2009a, p. 6-203).
      32 However, as noted in the ISA, in a sensitivity analysis when selectively removing cities from the overall estimate,
      the significant association between the PM10 mortality risk estimate and the PM2 5 Ni fraction was diminished upon
      removing New York City from the analysis, which is consistent with the results presented by Dominici et al. (2007)
      (US EPA, 2009a, section 6.5.2.5; Figure 6-32).
      33 Of note, New York City was not included in the 25 cities examined by Franklin et al. (2008).
             Draft Do Not Quote or Cite           2-62                         March 2010

-------
 1         Limited evidence is available to evaluate the health effects associated with long-term
 2    exposures to PM2.5 components (US EPA, 2009a, section 7.6.2). The most significant new
 3    evidence is provided by a study that evaluated multiple PM2 5 components and an indicator of
 4    traffic density in an assessment of health effects related to long-term exposure to PM2.5 (Lipfert
 5    et al., 2006). Using health data from a cohort of U.S. military veterans and PM2.5 data from
 6    EPA's CSN, Lipfert et al. (2006) reported positive associations between mortality and long-term
 7    exposures to nitrates, EC, Ni and V as well as traffic density and peak O3 concentrations.
 8    Additional evidence from a long-term exposure study conducted in a Dutch cohort provides
 9    supportive evidence that long-term exposure to traffic-related particles is associated with
10    increased mortality (Breelen et al., 2008).
11          With respect to source categories of fine particles associated with a range of health
12    endpoints, the ISA reports that currently available evidence suggests associations between
13    cardiovascular effects and a number of specific PM2.5 -related source categories, specifically oil
14    combustion,  wood or biomass burning, motor vehicle emissions, and crustal or road dust sources
15    (US EPA, 2009a, section 6.6; Table 6-18). In addition, a few studies have evaluated associations
16    between PM2.5-related source categories and mortality. These studies included a reported
17    association between mortality and a PM2.5 coal combustion factor (Laden et al., 2000), while
18    others linked mortality to a secondary SC>42 long-range transport PM2.5 source (Ito et al., 2006;
19    Mar et al., 2006) (US EPA, 2009a, section 6.6.2.1). There is less consistency in associations
20    observed between PM2.5  sources and respiratory health effects, which may be partially due to the
21    fact that fewer studies have been conducted that evaluated respiratory-related outcomes and
22    measures.  However, there is some evidence for associations with secondary SC>42~.and
23    decrements in lung function in asthmatic and healthy adults (US EPA, 2009a, p. 6-211; Gong et
24    al., 2005; Lanki et al., 2006). Respiratory effects relating to the crustal/soil/road dust and traffic
25    sources of PM have been observed in asthmatic children and adults (US EPA, 2009a, p. 6-205;
26    Gent et al., 2009; Penttinen et al., 2006).
27          Recent studies have shown that source apportionment methods have the potential to add
28    useful insights into which sources and/or PM constituents may contribute to different health
29    effects.  Of particular interest are several epidemiologic studies that compared source
30    apportionment methods and reported consistent results across research groups (US EPA, 2009a,
31    p. 6-211; Hopke et al.,  2006; Ito et al., 2006; Mar et al., 2006; Thurston et al., 2005).  These
32    studies reported associations between total mortality and secondary  sulfate in two cities for two
33    different lag times. The  sulfate effect was stronger for total mortality in Washington D.C. and
34    for cardiovascular-related morality in Phoenix (US  EPA, 2009a, p. 6-204).  These studies also
35    found some evidence for associations with mortality and a number of source categories (e.g.,
36    biomass/wood combustion, traffic, copper smelter,  coal combustion, sea salt) at various lag times
            Draft Do Not Quote or Cite          2-63                         March 2010

-------
 1    (US EPA, 2009a, p. 6-204). Sarnat et al., (2008) compared three different source apportionment
 2    methods and reported consistent associations between ED visits for cardiovascular diseases with
 3    mobile sources and biomass combustion as well as increased respiratory-related ED visits
 4    associated with secondary sulfate (US EPA, 2009a, pp. 6-204 and 6-211; Sarnat et al., 2008).
 5           In summary, in considering the currently available evidence for health effects associated
 6    with chemical components and source categories of PM2.5 as presented in the ISA, we reach the
 7    preliminary conclusion that additional information available in this review continues to provide
 8    evidence that many different constituents of the fine particle mixture as well as specific source
 9    categories of fine particles are linked to adverse health effects. However, as noted in the ISA,
10    while "[tjhere is some evidence for trends and patterns that link particular ambient PM
11    constituents or sources with specific health outcomes.. .there is insufficient evidence to
12    determine whether these patterns are consistent or robust" (US EPA, 2009a, p. 6-210).
13    Furthermore, the ISA concludes that "the evidence is not yet sufficient to allow differentiation of
14    those constituents or sources that are more closely related to specific health outcomes" (US EPA,
15    2009a,  pp. 2-26 and 6-212). Therefore, we reach the preliminary conclusion that the currently
16    available evidence is not sufficient to support consideration of a separate indicator for a specific
17    PM2.5 component or source category of fine particles.  Furthermore, we also reach the
18    preliminary conclusion that the evidence is not sufficient to support eliminating any component
19    or source from the mix of fine particles included in the PM2 5 indicator.
20           We recognize that much research supported by EPA, the Health Effects Institute (HEI),
21    the Electric Power Research Institute (EPRI), and  others is underway to evaluate the role of
22    PM2.s components/sources and agree that additional research is needed to improve future
23    understanding of the role of specific fine particle components and/or sources of fine particles.
24    Summary
25           In considering whether currently available  evidence provides support for retaining,
26    revising, or supplementing the current PM2 5 mass-based indicator, we first reach the preliminary
27    conclusion that it is appropriate to retain PM2 5 as the indicator for fine particles. Secondly, we
28    reach the preliminary conclusion that the currently available evidence does not provide a
29    sufficient basis for supplementing the mass-based  PM2 5 indicator by considering a separate
30    indicator for ultrafme particles as a subtraction of  fine particles. Furthermore, we also reach the
31    preliminary conclusion that the currently available evidence is too limited to provide support for
32    considering a separate indicator for a specific PM2.s component or source category of fine
33    particles or for eliminating any individual component or source category from the mix of fine
34    particles included  in the PM2 5 mass-based indicator.
            Draft Do Not Quote or Cite           2-64                        March 2010

-------
 1    2.3.2  Averaging Times
 2          In the last review, EPA recognized that the available information related to exposure
 3    periods of concern was generally consistent and supportive of the conclusions reached in the
 4    review completed in 1997.  In that review, the Agency retained two PM2.5 standards, based on
 5    annual and 24-hour averaging times in order to provide protection for health effects associated
 6    with short- and long-term exposure periods (71 FR 61164, October 17, 2006). In this review, in
 7    evaluating alternative fine particle standards, we first address the element of the standard related
 8    to averaging time by asking the following question:
 9    •  To what extent does the currently available information continue to provide support for
10       the current 24-hour and annual averaging times?
11          In considering whether the information available in this review supports consideration of
12    different averaging times for PM2.5 standards, we note that the available information is generally
13    consistent with and supportive of the conclusions reached in the previous reviews to set a suite of
14    PM2.s  standards including standards with both annual and 24-hour averaging times. The
15    overwhelming majority of studies conducted since the last review continue to utilize 24-hour and
16    annual averaging times, and largely contribute to the body of evidence for health effects related
17    to both short-term (from less than 1  day to up to several days) and long-term (from a year to
18    several years) measures of PM2 5.  Consequently, our preliminary conclusion is that the  currently
19    available evidence continues to provide support for a 24-hour and annual averaging time.
20          With respect to a standard with an annual averaging time, we recognize that an annual
21    standard would provide effective protection against both annual and multi-year, cumulative
22    exposures that are associated with an array of health  effects. With regard to providing protection
23    for short-term fine particle exposures, we note that the large majority of short-term
24    epidemiologic studies report associations based on 24-hour averaging times or on multiple-day
25    averages (i.e., distributed lag). Furthermore, we recognize that a 24-hour standard can
26    effectively protect against episodes lasting several days, as  well as providing some degree of
27    protection from potential effects associated with shorter duration peak levels of PM2.5, and from
28    episodes that result in localized or seasonal PM2 5 exposures of concern in areas where the
29    highest 24-hour-to-annual mean PM2.5 ratios are appreciably above the national  average. In
30    consideration of the currently available evidence, we have focused on evaluating alternative
31    standards in the quantitative risk assessment conducted for  this review that retained the
32    averaging times of the current standards  (i.e, 24-hour and annual averaging times) (US EPA,
33    2010a)
34          In summary, we reach the preliminary conclusion that there is strong support in this
35    review for retaining the current 24-hour and annual averaging times.  We have then considered if

            Draft Do Not Quote or Cite           2-65                        March 2010

-------
 1    the currently available evidence provides support for supplementing the current averaging times
 2    with additional standards focused on subdaily, multi-day, or seasonal exposures by posing two
 3    more specific questions.
 4    •   To what extent does the currently available scientific evidence provide support for
 5       considering a standard with an averaging time less than 24 hours to address health
 6       effects associated with subdaily fine particle exposures?
 7          In the last review, we recognized the availability of limited evidence of effects associated
 8    with exposure periods shorter than 24-hours (e.g., one to several hours) (US EPA, 2004, section
 9    3.5.5.1) and concluded that, while this evidence was "too limited to serve as the basis for
10    establishing a shorter-than-24-hour fine particle primary standard," it provided "added weight to
11    the importance of a standard with a 24-hour averaging time" (71 FR 61164/2; US EPA, 2005,
12    section 5.3.3).  More specifically, the AQCD noted that while few epidemiological studies had
13    used ambient PM concentrations averaged over time intervals shorter than 24 hours,  several
14    epidemiological studies reported statistically significant associations between 2- to 4-hour PM2 5
15    concentrations  and cardiovascular health endpoints, including myocardial infarction (MI)
16    incidence and heart rate variability (HRV) (US EPA 2004, pp.  8-162 to 8-165). In particular,
17    Peters et al (2001) reported effect estimates for MI incidence with PM2.5 averaged over 2- and
18    24- hours that were similar in magnitude and statistically significant (US EPA 2004,  p. 8-165).
19          In this review, a much larger body of more recent studies provide additional evidence of
20    cardiovascular  effects associated with  exposure periods shorter than 24 hours (US EPA, 2009a,
21    section 6.2).  In this review, two studies that assessed sub-daily and daily exposures did not
22    observe an association between PM2 5  and risk of MI, however no association was evident with
23    24-hour exposures to PM2.5 as well (US EPA, 2009a, p. 6-67; Sullivan et al.,  2005; Peters et al.,
24    2005). One study found a strong positive association between  self-reported exposure to traffic
25    and the onset of an MI within one hour, although this study did not directly measure traffic-
26    related pollution (US EPA 2009a, p. 6-67;  Peters et al. 2004).  For cardiac arrests witnessed by
27    bystanders, another study found a significant association with PM2 5 exposure during the hour of
28    the arrest, with even larger risk estimates for older adults (ages 60-75) or those that presented
29    with asystole, a particular form of cardiac arrest when electrical activity in the heart stops (US
30    EPA 2009a, p.  6-77; Rosenthal et al. 2008).
31          With respect to heart rate and HRV, epidemiological studies reported effects associated
32    with pollutant concentrations lagged as short as 1 to 2 hours, but more consistently with lags of
33    24 to 48 hours. The results of several  new controlled human exposure studies provide limited
34    evidence to suggest that acute exposures (2-hour) to near ambient levels of concentrated ambient
35    particles, or CAPs (PM2 5, PMio-2.s and ultrafme particles) may be associated with small changes
36    in HRV (USEPA, 2009a, p. 6-10). Changes in HRV parameters, however, are variable with
            Draft Do Not Quote or Cite           2-66                         March 2010

-------
 1    some showing increased parasympathetic activity relative to sympathetic activity and others
 2    showing the opposite.  One study reported adverse associations of 2-hour exposures to PMio with
 3    implantable cardiac defibrillator-detected ventricular arrythmias (US EPA, 2009a, p. 6-15;
 4    Ljungman et al., 2008).  Two panel studies of elderly subjects found electrocardiogram changes
 5    typically representative of cardiac ischemia (S-T segment depression) significantly associated
 6    with exposures to PM2.5 of less than 24 hours (US EPA 2009a, p. 6-21; Gold et al., 2005; Lanki
 7    et al. 2008).
 8           With respect to vasomotor function, the systemic vasculature is likely to be a target
 9    organ.  Endothelial dysfunction is a factor in many diseases and may contribute to the origin
10    and/or exacerbation of perfusion-limited diseases, such as MI or ischemic heart disease (IHD), as
11    well as hypertension. Endothelial dysfunction is also a characteristic feature of early and
12    advanced atherosclerosis. Six studies found effects of exposures of 2 hours or less on other
13    vasomotor outcomes.34 The evidence on blood pressure changes is more mixed, with one panel
14    study (Dales et al.,  2007; EPA 2009, p. 6-35) finding no association with blood pressure changes
15    with 2-hour exposures to traffic-related PM2.5; while one panel and two controlled human
16    exposure studies (US EPA, p. 6-37; Chuang et al., 2005; Urch et al., 2005 and Fakhri et al.,
17    2009) reported changes in blood pressure from  1- to 3-hour exposures to submicrometer
18    particles, and 2-hour exposures to a combination of Os and CAPs, respectively.
19           One panel study that investigated associations between ambient PM and markers of
20    systemic inflammation  among senior citizens (> 60 years of age) (US EPA, 2009a, p. 6-41;
21    Dubowsky  et al., 2006), found positive associations between daily PM2.5 levels and measures of
22    C-reactive protein (CRP), interleukin-6 (IL-6), and white blood cells (WBCs). When the
23    analysis was limited to  exposures that occurred only on daily bus trips (2 hours), the effect
24    estimates were similar in direction to the main analysis but were smaller in magnitude, with the
25    authors expressing  the view that this was likely due to greater measurement error. Another panel
26    study assessed the effects of in-vehicle exposure to PM2.5 of healthy young non-smoking male
27    state troopers working from 3 PM to midnight (US EPA, 2009a, p. 6-41 and 6-48; Riediker et al.,
28    2004).  This study concluded that PM2.5 originating from speed-changing traffic modulates the
29    autonomic control of the heart rhythm, increases the frequency of premature supraventricular
30    beats and elicits proinflammatory and pro-thrombotic responses in healthy young men. New
31    studies involving controlled human exposures to various particle types have provided limited and
32    inconsistent evidence of a PM-induced increase in markers of systemic inflammation (US EPA,
      34 Two studies (Dales et al. 2007; Rundell et al. 2007; US EPA 2009a, p. 6-25) found reductions in flow mediated
      dilatation (FMD) with 30-minute and 2-hour exposures to PM2 5, respectively. Four other studies (Peretz et al.,
      2008; Lund et al., 2009; Rundell and Caviston, 2008; and Shah et al., 2008; US EPA 2009a, p. 6-28 to 6-29) found
      effects of exposures of 2 hours or less on other vasomotor outcomes, including brachial artery diameter (BAD),
      circulating levels of endothelin-1, and reactive hyperemia, respectively.
            Draft Do Not Quote or Cite          2-67                        March 2010

-------
 1    2009a, p. 6-46). New findings for hemostasis, thrombosis and coagulation from controlled
 2    human exposure studies have provided in consistent evidence that short-term exposure to PM at
 3    near ambient levels may have small, yet statistically significant effects on hemostatic markers in
 4    healthy subjects or people with coronary artery disease (US EPA, 2009a, p. 6-50).
 5          With respect to respiratory health outcomes, in the last review, evidence of association
 6    between respiratory symptoms and subdaily exposures to PMio was available from two US panel
 7    studies of symptoms in asthmatic subjects (US EPA 2004, section 8.3.3.1.1). These two studies
 8    used 1-hour and 24-hour average concentrations.  The PMio 1-hour outcome was larger than the
 9    24-hour outcome for lower respiratory illness in one study, but the reverse was true for cough in
10    the other study.
11            In this review, evidence of subdaily PM2 5 exposures  associated with respiratory diseases
12    continues to be limited. The ISA concludes that for several studies of hospital admissions or
13    medical visits for respiratory diseases, the strongest associations were observed with 24-hour
14    average or longer exposures, not with less then 24-hour exposures (US EPA, 2009a, section 6.3).
15    A study that examined evidence of respiratory effects associated with subdaily PM2.5 exposures
16    in two New York City communities by assessing the association between 24-hour and 1-hour
17    maximum PM2.5 levels and ED visits. This study reported an  increase in asthma-related ED
18    visits that was similar for the two time periods (New York State Department of Health, 2006).
19    These results were robust to adjustment for copollutants.  However, a panel study found an
20    association between a 10 |ig/m3 increase in morning maximum 1-hour mean, but not daily, PM2.5
21    concentrations and increased likelihood of rescue medication use in asthmatic children (US EPA
22    2009a, p. 6-90; Rabinovitch et al., 2006).  Another study examined the association between a
23    marker for pulmonary inflammation (exhaled nitric oxide, or eNO), and ambient PM2 5
24    concentrations. Each 10 |ig/m3 increase in 1-hour mean PM2 5 concentration was associated with
25    a statistically significant increase in eNO among asthmatic children not taking inhaled
26    corticosteroids (US EPA 2009, p. 6-101; Mar et al., 2005).
27          In summary, there is a rapidly growing body of studies that provide evidence for
28    cardiovascular effects associated with subdaily exposure to PM, especially effects related to
29    HRV, cardiac ischemia, and vasomotor function, and also of changes in markers of systemic
30    inflammation, hemostasis, thrombosis and coagulation. Because these studies have used
31    different indicators (e.g., PM2.5, PMio, PMio-2.5, ultrafme particles), averaging times, and health
32    outcomes, it is difficult to draw conclusions about cardiovascular effects associated specifically
33    with subdaily exposures to PM2.5. Although there is additional evidence of respiratory effects
34    associated with PM2.5 exposure periods shorter than 24-hours  (e.g., one to several hours), this
35    evidence is much sparser than for cardiovascular effects.  Considering the currently available
36    evidence, we reach the preliminary conclusion that this information is too unclear, with respect
            Draft Do Not Quote or Cite           2-68                        March 2010

-------
 1    to the indicator, averaging time and health outcome, to serve as a basis for establishing a shorter-
 2    than-24-hour PM2.5 primary standard at this time. However, this evidence does provide added
 3    weight to the importance of a standard with a 24-hour averaging time. We recognize that the
 4    assessment of health effects associated with shorter-than-24-hour exposure is an important area
 5    of research that could provide a basis for the consideration of a subdaily PM standard in the
 6    future.  We note that the availability of hourly PM2.5 concentrations from the PM2.5 continuous
 7    monitors reporting to the AQS should provide the basis for a clearer assessment of the effects of
 8    subdaily exposure to PM2.5 in the future.
 9    •  To what extent does the currently available scientific evidence provide support for
10       considering separate standards with distinct averaging times to address effects
11       associated with seasonal fine particle exposures?
12          With regard to health effects associated with PM2.5 exposure across varying seasons in
13    this review, Bell et al. (2008) reported higher PM2.5 risk estimates for hospitalization for
14    cardiovascular and respiratory diseases in the winter compared to other seasons in a large,
15    national multi-city study.  In comparison to the winter season, smaller statistically significant
16    associations were  also reported between PM2.5 and cardiovascular morbidity for spring and
17    autumn, and a positive non-significant association was observed for the summer months. In the
18    case of mortality, Zanobetti and  Schwartz (2009) reported in their multi-city US based study a 4-
19    fold higher effect estimate for PM2.5 associated mortality for the spring as compared to the
20    winter.  These results suggest individuals are at greater risk of dying from higher exposures to
21    PM2.5 in the warmer months, and at greater risk of PM associated hospitalization for
22    cardiovascular and respiratory diseases during colder months of the year.  Overall, we observe
23    that there are few studies presently available to deduce a general pattern in PM2.5 risk across
24    seasons. In addition, these studies utilized 24-hour exposure periods  within each season to assess
25    the PM2.5 associated health effects, and do not provide information on health effects associated
26    with a season-long exposure to PM2.5.  Due to these limitations in the currently available
27    evidence, we reach the preliminary conclusion that there is no basis to consider a seasonal
28    averaging time separate from a 24-hour averaging time.
29    Summary
30    In summary, we recognize that the currently available evidence informs our understanding of
31    exposure durations of concern and continues to provide strong support for standards that provide
32    protection for both short- and long-term exposures. In considering the possibility of effects
33    associated with subdaily PM2.5 exposures (i.e., less than 24-hour exposures) we recognize that
34    there is additional evidence available in this review, primarily focused on cardiovascular effects
35    with more limited evidence for respiratory effects.  However, because these studies have used
36    different indicators of PM exposure (e.g., PM2.5, PMio, UFPs), averaging times, and a broad
            Draft Do Not Quote or Cite          2-69                        March 2010

-------
 1    range of health outcomes, it is difficult to use this evidence to serve as a basis for establishing a
 2    standard with a shorter-than-24-hour averaging time.  With respect to seasonal effects, while we
 3    recognize there is some new evidence  for PM2 5. related effects differentiated by season, we reach
 4    the preliminary conclusion that this evidence is, at this point, too limited to use as a basis for
 5    establishing a PM2.5 standard with a seasonal averaging time. Based on the above considerations,
 6    we initially conclude that the currently available information provides strong support for
 7    retaining the current annual and 24-hour averaging times but does not provide  support for
 8    alternative averaging times of less than 24-hours or for seasons

 9    2.3.3   Forms
10           The "form" of a standard defines the air quality statistic that is to be compared to the
11    level of the standard in determining whether an area attains the standard. In this review, staff are
12    evaluating whether currently available information support consideration of alternative forms for
13    the annual or 24-hour PM2.5 standards.

14           2.3.3.1 Form of the Annual Standard
15           In 1997, EPA established the form of the annual PM2 5 standard as an annual arithmetic
16    mean, averaged over 3 years, from single or multiple community-oriented monitors.  This form
17    was intended to represent a relatively stable measure of air quality and to characterize area-wide
18    PM2.5 concentrations.  The arithmetic mean served to represent the broad distribution of daily air
19    quality values, and a 3-year average provided a more stable risk reduction target than a single-
20    year annual average. When setting the initial fine particle standards in 1997, the level of the
21    annual PM2.5 standard was to be compared to measurements made at the community-oriented
22    monitoring site recording the highest level, or, if specific constraints were met35, measurements
23    from multiple community-oriented monitoring sites could be averaged (62 FR 38671 to 38672,
24    July 18, 1997). The constraints for allowing the use of spatially averaged measurements were
25    intended to limit averaging across poorly correlated or widely disparate air  quality values to
26    ensure that spatial averaging would not result in inequities in the level of protection provided by
27    the PM2.5 standards against health effects associated with short- and long-term PM2.5 exposures
28    (62 FR 38672). This approach was judged to be consistent with the epidemiologic studies on
29    which the PM2.5 standard was primarily based, in which air quality data were generally averaged
30    across multiple monitors in an area or  were taken from a single monitor that was selected to
31    represent community-wide exposures,  not localized "hot spots."
      35 The original criteria for spatial averaging included: (1) the annual mean concentration at each site shall be within
      20 percent of the spatially averaged annual mean, and (2) the daily values for each monitoring site pair shall yield a
      correlation coefficient of at least 0.6 for each calendar quarter.
             Draft Do Not Quote or Cite           2-70                        March 2010

-------
 1            In the last review, in considering the form of the primary annual PM2.5 standard, EPA
 2    posed the question as to whether an annual standard that allowed for spatial averaging, within the
 3    original specified or alternative constraints, would provide appropriate public health protection.
 4    Analyses conducted in the last review, based on a much larger set of PM2.5 air quality data than
 5    was available for the review completed in 1997, assessed both aggregate population risk across
 6    an entire urban area and the potential for disproportionate impacts on potentially vulnerable
 7    populations within an area.  The results of these analyses provided evidence concerning the
 8    potential for disproportionate impacts on potentially vulnerable populations, noting "the highest
 9    concentrations in an area tend to be measured at monitors located in areas where the surrounding
10    population is more likely to have lower education and income levels, and higher percentages of
11    minority populations" (71 FR 61166/2, see also US EPA, 2005,  section 5.3.6.1; Schmidt et al.,
12    2005, Attachment A/Analysis 7).36
13           In addition,  the effect of allowing the use of spatial averaging on aggregate population
14    risk was considered as part of the sensitivity analyses included in the health risk assessment
15    conducted for the last review (US EPA, 2005, section 4.2.2).  In that analysis, changing the basis
16    of the annual standard design value from the concentration at the highest monitor to the average
17    concentration across all monitors reduced the air quality adjustment needed to just meet the
18    current or alternative annual standards.  As expected, the estimated risks remaining upon
19    attainment of the current annual standard in areas where the annual standards was the
20    "controlling standard" were greater when spatial averaging was used than when the highest
21    monitor was used (i.e., the estimated reductions in risk associated with just attaining the current
22    or alternative annual standards are less when spatial averaging was used). The estimated
23    mortality incidence associated with long-term exposure in most cases ranged from about 10 to
24    60% higher when spatial averaging was used, and estimated mortality incidence associated with
25    short-term  exposure in most cases ranged from about 5  to 25% higher.  In light of these analyses,
26    EPA retained the form of the standard as an annual arithmetic mean, averaged over 3 years with
27    modifications that strengthened the standard by tightening the criteria for use of spatial
28    averaging37 to provide increased protection for vulnerable populations exposed PM2.5.
      36 As summarized in footnote 29 at 71 FR 61166/2, the 2004 AQCD noted that some epidemiologic studies, most
      notably the ACS study of associations between long-term PM2 5 exposure and mortality, reported larger effect
      estimates in the cohort subgroup with lower education levels (US EPA, 2004, p 8-103). The 2004 AQCD also noted
      that lower education level may be a marker for lower socioeconomic status (SES) that may be related to increased
      vulnerability to the effects of fine particle exposures, for example, as a result of greater exposure from proximity to
      sources such as roadways and industry, as well as other factors such as poorer health status and access to health care
      (US EPA, 2004, section 9.2.4.5).
      37 The current criteria for spatial averaging include:  (1) the annual mean concentration at each site shall be within 10
      percent of the spatially averaged annual mean, and (2) the daily values for each monitoring site pair shall yield a
      correlation coefficient of at least 0.9 for each calendar quarter (71 FR 61167/2-3, October 17, 2006).
             Draft Do Not Quote or  Cite           2-71                          March 2010

-------
 1           In this review, the currently available evidence provides stronger support that
 2    socioeconomic status (SES), measured using surrogates such as educational attainment,
 3    residential location, and income level, modifies the association between PM and morbidity and
 4    mortality outcomes.  Thus, as discussed in section 2.2.1, the ISA concludes that susceptible
 5    populations include persons with lower SES levels.  In light of this information and the analyses
 6    conducted for the last review as discussed above, we believe that the existing constraints on
 7    spatial averaging, as modified in 2006, may not be adequate to avoid substantially greater
 8    exposures in some areas, potentially resulting in disproportionate impacts on persons with lower
 9    SES levels.  Therefore, we are reconsidering the appropriateness of continuing to allow spatial
10    averaging across monitors as part of the form of the annual PM2.5 standard. Recognizing that it is
11    the link between the form and the level of a standard that determines the degree of public health
12    protection the standard affords, we discuss consideration for eliminating the spatial averaging
13    provisions from the form of the annual PM2.5 standard in conjunction with consideration of
14    alternative levels to address health effects related to long-term PM2.5 exposures in section 2.3.4
15    below.

16           2.3.3.2 Form of the 24-Hour Standard
17           In  1997, EPA established the form of the 24-hour PM2.5 standard as the 98th percentile of
18    24-hour concentrations at each population-oriented monitor within an area, averaged over three
19    years (62 FR at 38671 to 38674, July 18, 1997). In making that decision, EPA recognized that a
20    concentration-based form gave proportionally greater weight to days when concentrations were
21    well above the level of the standard than to days when the concentrations were just above the
22    standard. Further, a concentration-based form better compensated for missing data and less-
23    than-every-day monitoring; and, when averaged over 3 years, it had greater stability and, thus,
24    facilitated the development of more stable implementation programs.  The Agency selected the
25    98th percentile as an appropriate balance between adequately limiting the occurrence of peak
26    concentrations and providing increased stability and robustness. In addition, by basing the form
27    of the standard on concentrations measured at population-oriented monitoring sites, EPA
28    intended to provide protection for people residing in or near localized areas of elevated
29    concentrations.
30           In the last review,  in conjunction with considering alternative 24-hour standard levels,
31    EPA concluded it was appropriate to retain a concentration-based form that was defined in terms
32    of a specific percentile of the distribution of 24-hour PM2.5 concentrations at each population-
33    oriented monitor within an area, averaged over 3 years. In that review, in considering retaining
34    the 98th percentile form or revising the standard to a 99th percentile from, the Agency evaluated
35    the combination of form and level  to provide appropriate public health protection and, in
             Draft Do Not Quote or Cite          2-72                        March 2010

-------
 1    particular, considered:  (1) the relative risk reduction afforded by alternative forms at the same
 2    standard level, (2) the relative year-to-year stability of the air quality statistic to be used as the
 3    basis for the form of a standard, and (3) the implications from a public health communication
 4    perspective of the extent to which either form allows different numbers of days in a year to be
 5    above the level of the standard in areas that attain the standard.  Based on these considerations,
 6    the Administrator concluded it was appropriate to retain the 98th percentile form of the 24-hour
 7    standard in conjunction with lowering the level of the standard. In reaching this final decision,
 8    EPA recognized a technical problem associated with a potential bias in the method used to
 9    calculate the 98th percentile concentration for this form.  As such, EPA adjusted the sampling
10    frequency requirement in order to reduce this bias. Accordingly, the Agency modified the final
11    monitoring requirements such that areas that are within 5 percent of the standards are required to
12    increase the sampling frequency to  every day (71 FR 61164 to 61165, October 17, 2006).
13           In this first draft PA, we have focused consideration of alternative 24-hour PM2.5
14    standards on alternative levels only, as discussed in section 2.3.5.  In the second draft PA, we
15    intend to explore more fully information regarding peak air quality concentrations to better
16    inform our understanding of the implications of retaining or modifying the current form of the
17    24-hour standard.

18    2.3.4   Alternative Levels to Address Health Effects Related to Long-term PMi.s Exposures
19           In considering alternative PM2.5 standards that would provide protection against health
20    effects related to long-term exposures, we have taken into account both evidence-based and risk-
21    based considerations. As discussed below, we first evaluate the available evidence from long-
22    term PM2.5  exposure studies, as well as the uncertainties and limitations in that evidence as
23    presented in the ISA, to assess the degree to which alternative annual PM2 5 standards can be
24    expected to provide  protection against effects related to long-term exposures (section 2.3.4.1).
25    Secondly, we have considered the quantitative risk estimates associated with long-term PM2.5
26    exposure, as discussed in the second draft RA, to assess the extent to which alternative standards
27    can be  expected to reduce the estimated risks attributable to long-term exposure to PM2.5 (section
28    2.3.4.2). With respect to the suite of PM2.5 standards, our preliminary integrated conclusions
29    presented in section 2.3.6 are based in part on the conclusions from this section and in part on
30    preliminary staff conclusions from  section 2.3.5, in which alternative PM2.5 standards to address
31    health effects related to short-term PM2.5 exposures are discussed.

32           2.3.4.1 Evidence-based Considerations
33           In taking into account evidence-based considerations for informing our understanding of
34    alternative levels to address health effects related to long-term fine particle exposures, we have

             Draft Do Not Quote or Cite           2-73                        March 2010

-------
 1    initially focused on long-term PM2.5 exposure studies conducted in the U.S. and Canada and
 2    placed the greatest weight on associations that have been judged in the ISA to be causal or likely
 3    causal. We have also considered the evidence for a broader range of health outcomes judged in
 4    the ISA to have suggestive evidence  of a causal association or that focus on specific susceptible
 5    populations to evaluate whether this evidence provides support for considering lower alternative
 6    levels. Collectively, we have integrated the currently available evidence to address the following
 7    question:
 8    •  To what extent does the currently available evidence provide support for revising the
 9       current suite of standards to provide  protection for long-term fine particle exposures?
10           In looking first at evidence from long-term PM2.5 exposure studies of mortality (causal
11    association), we note, as discussed in section 2.2.1, that the extended follow-up analyses of the
12    ACS and Harvard Six Cities studies,  have confirmed and strengthened evidence of associations
13    reported in the last review at lower air quality levels. In addition, new cohort studies, including
14    the WHI evaluating cardiovascular-related mortality in postmenopausal women and analyses of
15    mortality in a Medicare cohort, provide further evidence of effects associated with long-term
16    PM2.5 exposures at air quality levels in the same range as the more recent years of air quality data
17    evaluated in the ACS and Harvard Six Cities extended follow-up studies.
18           As outlined in Figure 2-1 long-term mean PM2.5 concentrations across all cities evaluated
19    in each of these studies range from 10.7 to 16.4 |ig/m3.38 In looking first at the long-term
20    exposure mortality studies, we note that the long-term mean PM2 5 concentration in the extended
21    follow-up of the Harvard Six Cities study was 16.4 |ig/m3. As noted in  section 2.2.1, in focusing
22    on the extended follow-up period (1990-1998), we estimate an aggregate long-term mean PM2 5
23    concentration across the six cities of 14.8 |ig/m3 (Laden, 2009).  Eftim et al., (2008) evaluated
24    mortality in a Medicare cohort consisting of adults age 65 and older, a susceptible population,
25    within the same six cities using more recent air quality with a long-term mean of 14.1 |ig/m3.  In
26    the extended ACS study, the mean for the more recent time period used in the analysis (1999 to
27    2000) was 14.0 |ig/m3. In looking at the association based on the air quality for both time
28    periods, the long-term mean PM2.s concentration was 17.1  |ig/m3 (Pope et al., 2004). Analysis of
29    mortality associated with long-term PM2.5 exposure using a Medicare cohort and the ACS
30    locations using more recent air quality reported a long-term mean PM2 5 concentration of 13.6
31    Hg/m3 (Eftim et al., 2008).  Additional evidence of mortality associated with  long-term PM2 5
32    exposure in the MCAPS study across a larger number of counties (668 vs. 110 counties) using
33    more extensive air quality data, provides evidence of effects in older adults, a susceptible
34    population, at similar levels.  This study included effect estimates for three geographic areas (i.e.,
      38 We note that Figure 2-1 also includes studies with substantially higher long-term mean PM2 5 concentrations.
            Draft Do Not Quote or Cite          2-74                        March 2010

-------
1
2
                  Figure 2-2.  Distribution of study-specific area mean PM2.s concentrations.
                              (a) Representative Long-Term Exposure Studies
                                                                 (b) Representative Short-Term Exposure Studies
                     30 -
                     25 -
                     20 -
                 o
                 o
                 U
10 -
                      5 -
                      o H

                           Study  EPA
                           Author  CM
                  EPA
                  max
                          Eftim-ACS
 Study  EPA
Author  CM
EPA
max
 Study  EPA
Author  CM
EPA
max
 Study  EPA
Author  CM
EPA
max
 Study  EPA   EPA
Author  CM   max
                                             Miller
             Bell
                                     Dominici06
                                     Zanobetti
                 Left plots represent the study author results; middle plots represent EPA replication results using a community monitor, CM, (spatial average) approach; and
                 right plots represent EPA replication results using a maximum value approach (max monitor mean for each year averaged over the study timeframe). Black
                 outlined boxes denote inter-quartile range and medians, stars denote means with cyan rectangles spanning +/- one standard deviation from the mean, and dots
                 represent the minimum and maximum values.
              Draft Do Not Quote or Cite
                                                              2-75
                     March 2010

-------
 1    eastern, central, western) with long-term median PM2.5 concentrations ranging from 10.7 to 14.0
 2    ng/m3 (Zeger et al., 2008).  In evaluating the air quality data considered in this study, we believe
 3    it is most appropriate to focus on the long-term median concentration across all cities which was
 4    reported as 13.2 |ig/m3 (Zeger et al., 2008).  A cystic fibrosis cohort study of mortality reported a
 5    positive but not statistically significant mortality association with an aggregate long-term mean
 6    PM2.5 concentration of 13.7 |ig/m3 in a susceptible population (Goss et al., 2004)
 7           In considering new evidence for cardiovascular effects associated with long-term PM2.5
 8    exposure (causal association), we note the WHI reported positive and statistically significant
 9    effects at a long-term average across cities of 13.5 |ig/m3 (Miller et al, 2007). As noted above in
10    section 2.2.1, this study considered the associations between long-term exposure to PM2.5 and
11    cardiovascular-related mortality as well as cardiovascular morbidity in post-menopausal women
12    with no previous history of cardiac disease.
13           Furthermore, with respect to respiratory effects associated with long-term PM2.5 exposure
14    (likely causal association), the continued follow-up of the Southern California CHS's cohort
15    study provides stronger evidence of decreased lung function growth in children, a susceptible
16    population, that persisted into early adulthood at a long-term mean PM2.5 concentration across 12
17    communities of 13.8 |ig/m3. McConnell et al., (2003) also reported a positive and statistically
18    significant effect between long-term PM2.5 concentrations and bronchitic symptoms as part of the
19    Southern California CHS.  These results are supported by a single-city cross-sectional study of
20    bronchitic symptoms in school-aged children reported a positive and statistically significant
21    association with long-term PM2.5 concentrations of 12 |ig/m3 (Kim et al., 2004).
22           In considering alternative levels for an annual standard that would provide protection
23    with an adequate margin of safety, we believe it is also appropriate to take into account evidence
24    of effects for which the reported associations provide only suggestive evidence of a causal
25    association, including, but not limited to, evidence of reproductive and developmental effects.
26    With respect to emerging evidence on low birth weight and infant mortality, especially related to
27    respiratory causes during the post-neonatal period, the ISA concludes that effects become "more
28    precise and consistently positive in locations with mean PM2.5 concentrations of 15 |ig/m3 and
29    above" (US EPA, 2009a, p. 2-13; section 7.4).  In particular, we note a recent study of
30    postneonatal infant mortality which observed a positive and statistically significant association
31    with an aggregate long-term mean PM2.5 concentration of 14.8 |ig/m3 (Woodruff et al., 2008).
32    We recognize that these effects are serious in nature and that data are continuing to emerge
33    related to reproductive and developmental outcomes.  At this time, however, the PM2.5
34    concentrations reported in studies evaluating these effects report ambient levels that are equal to
35    or greater than ambient levels observed in studies reporting mortality and cardiovascular and
36    respiratory effects. Therefore, in selecting alternative levels, we note that in providing protection
             Draft Do Not Quote or Cite           2-76                         March 2010

-------
 1    for mortality and cardiovascular and respiratory effects, it is reasonable to anticipate that
 2    protection will also be provided for reproductive and developmental effects.
 3           With respect to characterizing the concentration-response relationship for mortality
 4    associated with long-term PM2.5 exposures, we first note,  that in the last review, evidence was
 5    available to support a linear relationship. Using data from the ACS cohort, Pope et al., (2002)
 6    reported that the associations for all-cause, cardiovascular, and lung cancer mortality "were not
 7    significantly different from linear." In that study, the confidence intervals began to expand
 8    significantly starting around 13 to 12 |ig/m3, indicating greater uncertainty about the shape of the
 9    reported concentration-response relationship at and below this level (US EPA, 2004, Figure 8-7;
10    US EPA 2005, Figure3-4). In this review, additional evidence supports a linear concentration-
11    response relationship.  Schwartz et al. (2008) used a variety of statistical methods and reported
12    "the concentration-response curve was linear, clearly continuing below the level of the current
13    U.S. air quality standard of 15 |ig/m3" (US EPA, 2009a, p. 7-92).  No new evidence is available
14    in this review to inform our understanding of the confidence intervals around the estimated
15    concentration-response functions.
16           In considering what alternative levels would provide protection for effects observed in
17    such long-term PM2.5 exposure studies, we first recognize that the ISA has concluded that no
18    discernable threshold for an effect associated with long-term PM2.5 exposure can be identified
19    based on the currently  available evidence (US EPA 2009a, section 2.4.3). We recognize that
20    health effects may occur over the full range of concentrations observed in the studies, however,
21    we believe it is reasonable to conclude the evidence of association is strongest down to
22    somewhat below the aggregated mean concentration, such as down to one standard deviation
23    below the mean or to the lower end of the interquartile range, which includes the range in which
24    the data in the study are most concentrated. This approach is consistent with considering the
25    serious nature of the observed effects, including, but not limited to, mortality and cardiovascular
26    effects for which there is strong evidence of a causal association as well as respiratory effects,
27    for which there is strong evidence of a likely causal relationship. It also includes consideration of
28    the range of long-term average concentrations across cities and the point where the data become
29    less robust, suggestive of a concentration below which the association becomes appreciably more
30    uncertain. In evaluating the long-term exposure studies, we recognize that the overall density of
31    air quality concentrations becomes noticeable sparse below the range of air quality levels
32    represented by the lower interquartile range or one standard deviation below the mean. In our
33    view, an annual standard set below this range would be highly precautionary, giving little weight
34    to the remaining uncertainties in the broader body of evidence, including the limited number of
35    cities included in the long-term epidemiologic studies that reported long-term below the lower
36    interquartile range.
             Draft Do Not Quote or Cite          2-77                        March 2010

-------
 1           As discussed in section 1.2.1, this approach is relevant to inform judgments about
 2    providing an adequate margin of safety to prevent pollution levels that may pose an unacceptable
 3    risk of harm, even if the risk is not precisely identified as to nature or degree.  As presented in
 4    Figure 2-1, the lower interquartile range of long-term average concentrations or one standard
 5    deviation below study mean air quality levels as reported in the long-term mortality studies for
 6    the most recent years evaluated, report air quality concentrations in the range from about 10 to
 7    11 |ig/m3. With respect to considering air quality levels reported in the extended follow-up to
 8    the Southern California CHS, we recognize that using this approach, air quality levels one
 9    standard deviation below the mean across communities is significantly lower, about 6 |ig/m3, and
10    close to or below the LML reported in the majority of the long-term mortality studies.
11           In considering not only the level but also the form of the annual standard39, we
12    acknowledge that there may be differences between mean PM2.5 concentrations averaged across
13    monitors within a city/county as is typically considered in epidemiologic studies, compared to
14    the current form of the annual PM2.5 standard which typically focuses on ambient measurements
15    from the highest reporting community-oriented monitor. As such, we requested additional air
16    quality distribution data from authors of specific epidemiologic studies in order to effectively
17    place key epidemiologic studies in a policy-relevant context (Hassett-Sipple and Stanek, 2009).
18    Data were received for some but not all studies for which information was requested. To
19    broaden this data base, OAQPS conducted additional analyses for selected long-term PM2 5
20    exposure studies to better understand air quality distributions averaged across monitors in
21    comparison to air quality levels focused on the highest reporting monitors (Schmidt, 2010).
22    These analyses focused on selected long-term mortality studies and, as presented in Table 2-4
23    Figure 2-2a, reported composite monitor values  close to the long-term mean concentrations
24    reported by the study authors.40 As was expected, estimated air  quality levels for the highest
25    reported monitors, within a range of 14.9 |ig/m3 (Eftim et al., 2008) to 16.8 |ig/m3 (Miller et al,
26    2007), were higher than for the composite monitor values, within a range of 12.9 |ig/m3 (Miller
27    et al., 2007) to 14.0 |ig/m3 (Krewski et al., 2009),. The most notable difference was reported for
28    the WHI study which assessed only one year of air quality data (2000) for 36 study areas
      39 As summarized in section 2.3.3, the level of the annual PM25 standard is to be compared to measurements made at
      the community-oriented monitoring site recording the highest level, or if specific constraints are met, measurements
      from multiple community-oriented monitoring sites could be averaged ("spatial averaging").
      40 As noted by Schmidt (2010), differences between long-term mean concentrations reported by the study authors
      and the parallel values in the EPA analyses may be related to the EPA assessment only encompassing Federal
      Reference and Federal Equivalent Method air quality data while the study authors may have included additional
      (non-reference/equivalent) data. Also, there may be additional differences in the sets of monitoring sites utilized
      due to uncertainties associated with the geographical area boundary definitions.
             Draft Do Not Quote or Cite           2-78                        March 2010

-------
Table 2-4. Comparison of Air Quality Data for Selected Epidemiological Studies - Long-term Mean Concentrations
                                                                                                               41
Author Reported Data
Study
Cite
Years
Mean
(SD)
Mean
-1SD
Range
EPA Analysis (Schmidt 2010)
Composite Monitor
Mean
(SD)
Mean -
1SD
Lower
end of
IQR
range
(25%)
Range
Maximum Monitor
Mean
(SD)
Mean -
ISO
Lower
end of
IQR
range
(25%)
Range
Long-term Exposure Studies
ACS-
extended
follow-up
Medicare
-ACS
Women's
Health
Initiative
Krewski
et al 2009
Eftim et al
2008
Miller et
al 2007
1999-
2000
2000-
2002
2000
14.0
(3.0)
13.6
(2.8)
13.5
(3.3)
11.0
10.8
10.2
5.8-22
6-25
3.4-28
14.0
(3.5)
13.3
(3.1)
12.9
(3.0)
10.5
10.2
9.9
10.9
10.8
10.8
6-21
7-20
5-18
15.4
(4.1)
14.9
(3.6)
16.8
(4.3)
11.3
11.3
12.5
12.3
12.0
14.2
9-25
9-24
5-28
Short-term Exposure Studies
MCAPS
-Original
-Extended
National
Mortality
Study
Dominici
et al 2006
Bell et al
2008
Zanobetti
&
Schwartz,
2009
1999-
2002
1999-
2005
1999-
2005
13.4
(2.9)
12.9
(2.7)
13.2
(2.9)
10.5
10.2
10.3
11.3-
15.2
(IQR)*
4-20
7-25
13.3
(2.9)
12.9
(2.7)
13.7
(3.0)
10.4
10.2
10.7
11.2
11.0
11.6
5-21
4-19
6-26
13.9
(3.4)
13.4
(3.1)
14.1
(3.2)
10.5
10.3
10.9
11.6
11.4
12.0
5-29
5-26
7-26
41 All concentrations reported as ug/m3. * IQR = interquartile range




Draft Do Not Quote or Cite                           2-79
March 2010

-------
 1    (average across monitors and across cities (composite monitor) of 12.9 |ig/m3, compared to an
 2    average of the highest reporting monitors across cities of 16.8 |ig/m3).  When more study areas
 3    were considered and/or more years of air quality data were evaluated, the difference between the
 4    long-term mean for the composite monitor across cities compared to the long-term mean for the
 5    highest reporting monitor across cities was decreased.
 6           In considering the currently available evidence, we reach the preliminary conclusion that
 7    the long-term PM2.5 exposure studies provide a basis for considering alternative levels for the
 8    annual PM2.5 standard within a  range of about 13 |ig/m3 down to about 10 |ig/m3. A standard in
 9    the range of 13 to 12 |ig/m3 would reflect placing weight on setting a level somewhat below the
10    long-term mean PM2.5 concentrations evaluated in long-term PM2.5 exposure studies that show
11    associations with mortality and cardiovascular and respiratory morbidity over a lower range of
12    air quality levels than had been observed in the studies available in the last review. A standard in
13    the range of 11 to 10 |ig/m3 would be consistent with a judgment that, recognizing the serious
14    nature of the effects and that no discernable threshold for these effects can be identified,
15    appreciable weight should be accorded to considering the lower interquartile range of long-term
16    average concentrations within a study, or a range within one standard deviation around the study
17    mean of either the composite monitor or the highest reporting monitor.
18           Furthermore, recognizing that there is a link between the form and the level of a standard
19    that determines the degree of public health protection the standard affords, and in light of
20    stronger evidence available in this review identifying persons with lower SES levels as a
21    susceptible population as discussed in section 2.2.1, we recognize that the existing constraints on
22    spatial averaging,  as modified in 2006, may not be adequate to avoid substantially greater
23    exposures in some areas, potentially resulting in disproportionate impacts on persons with lower
24    SES levels.  Therefore, we reach the preliminary conclusion that it is appropriate to consider a
25    form of the annual PM2.5 standards that does not allow for the use of spatial averaging across
26    monitors such that the annual PM2.5 standard would be compared to measurements made at
27    monitoring sites that represent "community-wide air quality" recording the highest levels only.
28    We  plan to  conduct additional air quality analyses to explore this issue to inform the second draft
29    PA.

30           2.3.4.2 Risk-based Considerations
31           Beyond looking directly at the relevant epidemiologic evidence, staff have also
32    considered the extent to which  specific levels of alternative PM2.5 standards are likely to reduce
33    the estimated mortality risks attributed to long-term exposure to PM2.5 and the uncertainties in
34    the estimated risk reductions. We have based this evaluation on the results of the quantitative risk
35    assessment presented in the second draft RA (US EPA, 2010a) and posed a series of questions to
            Draft Do Not Quote or Cite           2-80                        March 2010

-------
 1    aid in considering how the current suite of PM2.5 standards might be revised to provide requisite
 2    public health protection.
 3    •   To what extent do alternative standards reduce estimated risks associated with long-
 4       term PM2.s exposure compared to risks associated with simulating air quality to just
 5       meet the current standards? What roles do the current 24-hour and annual standards
 6       have in simulating the risks remaining upon just meeting the alternative suites of
 7       standards considered in the quantitative risk assessment? What level of confidence is
 8       associated with these risk estimates?
 9          As noted in section 2.1.1, in summarizing risk estimates associated with long-term
10    exposure-related mortality, we focus on IHD-related mortality due to greater overall support for
11    a causal association with this endpoint.  However, we note that risk estimates were also
12    generated for all-cause, cardiopulmonary and lung cancer-related mortality. In presenting these
13    results, we focus first on the degree of risk reduction (and estimates of risk remaining) upon
14    simulation of the  alternative annual standard levels considered in the risk assessment. We then
15    discuss the nature and magnitude of risk reductions associated with alternative 24-hour standard
16    levels. Contrasting the degree of risk reduction provided by the suites of alternative annual and
17    24-hour standards is covered in the overall summary discussion presented at the end of this
18    section.
19    Risk reduction associated with alternative annual standards (14/35,13/35, and 12/35)
20          In evaluating alternative standards, we first focused on consideration of different levels
21    for the annual PM2.5 standard (e.g.,  14 |ig/m3, 13 |ig/m3, 12 |ig/m3) in conjunction with retaining
22    the current level of the 24-hour standard (35 |ig/m3).  We observed a consistent pattern of
23    increasing risk reduction with decreasing alternative annual standard levels, both in terms of the
24    number of urban study areas experiencing risk reductions and the magnitude of those reductions
25    (US EPA, 2010a, section 6.2.2).  Specifically, 5 of the 15 urban study areas experienced risk
26    reductions under the alternative annual standard level of 14 |ig/m3, with percent reductions in
27    PM2 s-attributable long-term exposure-related mortality (relative to risk under the current suite of
28    standards) ranging from 9%  (Baltimore) to 12% (Houston and Birmingham) (Figure 2-3  and US
29    EPA, 2010a, Appendix E, Table E-27).  For an annual standard level of 12 |ig/m3, 12 of the 15
30    urban study areas experience risk reductions, with percent reductions (relative to risk under the
31    current suite of standards) ranging from 11% (Phoenix) to 35% (Houston and Birmingham)
32    (Figure 2-3 and US EPA, 2010a, Appendix E, Table E-27). Note, that even in the 12/35  case,
33    three of the urban study areas (Tacoma, Fresno and Salt Lake City) did not experience any
34    decreases in risk,  although risk reductions were seen for these three study areas when alternative
35    24-hour standards were considered, as discussed below.
            Draft Do Not Quote or Cite           2-81                        March 2010

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
                                           Figure 2-3. Percent reduction in long-term exposure-related mortality risk
   (alternative standards relative to the current standards) (Note: inset shows PM2 5 related incidence and percent of total incidence for IHD mortality under the
                                                                current suite of standards)
                      60%
                   CO
                I  c
                ro  £
                75  *-
                3  2
                o  *=
                Q- c
                x  o
                UJ  S
                £  =
                5  •§
                51
                      40%
                      20%
                           Atlanta, GA 277 (227-324); 16.7% (13.7% -19.5%)
                           Baltimore, MD 374 (307-440); 14.7% (12.1% -17.3%)
                           Birmingham, AL 165 (135-194); 13.8%  (11.3% -16.2%)
                          • Dallas, TX 247 (202-291); 11.4% (9.3%-13.4%)
                           Detroit, Ml 478 (390-563); 11.5% (9.4%-13.5%)
                           Fresno, CA 98 (80-116); 8.5% (7%-10.1%)
                           Houston, TX 434  (355-511); 13.6% (11.1%-16%)
                           Los Angeles, CA 1094 (890-1296); 7.7% (6.3%-9.1%)
                           New York, NY 2222 (1814-2620); 11.8% (9.6%-13.9%)
                           Philadelphia, PA 330 (270-389);  13.2% (10.8% -15.6%)
                           Phoenix, AZ 402 (327-476); 8.5% (6.9%-10.1%)
                           Pittsburgh, PA 324 (264-382); 11.8% (9.6%-13.9%)
                           Salt Lake City, UT 19 (16-23); 3.7% (3%-4.4%)
                           St. Louis, MO 563 (461-662); 14.2% (11.6% -16.7%)
                           Tacoma, WA 49 (40-58); 4.7% (3.8%-5.6%)
                                  15/35**
                                              14/35
    13/35              12/35
Current and Alternative Standards
13/30
12/25
*Based on Krewski et al. (2009), exposure period from 1999 - 2000. The legend contains, for each urban area, the incidence estimate (and 95% CI) and the
estimate of percent of total incidence (and 95% CI) under the current standards.
**The current standards consist of an annual standard of 15 ug/m3 and a daily standard of 35 ug/m3. Combinations of an annual standard (n) and a daily standard
(m) are denoted n/m in this figure.
***The percent reductions for Salt Lake City and Tacoma at the 12/25 standard are 100% and 93%, respectively.
               Draft Do Not Quote or Cite
                                                                          2-82
                           March 2010

-------
 1           While there is a consistent pattern of risk reduction across the alternative annual
 2    standards with lower standard levels resulting in more urban study areas experiencing
 3    increasingly larger risk reductions, there is considerable variability in the magnitude of these
 4    reductions across urban study areas for a given alternative annual standard level (e.g., as noted
 5    above, for the alternative annual standard level of 12 |ig/m3, risk reduction ranged from 11% for
 6    Phoenix to 35% for Houston).  The variability in risk estimates reflects differing degrees of
 7    reduction in annual average PM2.5 concentrations across the study areas. The differences in
 8    annual average concentrations result, in part, because the study areas begin with varying annual-
 9    average PM2.5 concentrations after simulating just meeting the current suite of standards.
10    Therefore, even if study areas have similar "ending" annual average PM2.5, because the starting
11    point in the calculation (i.e., the annual-average PM2 5 concentrations upon just meeting the
12    current suite of standards) is often variable, the overall reduction in annual-average PM2.5
13    concentrations across the standards is also variable.42
14           The sensitivity analysis involving application of peak shaving rollback reveals that the
15    pattern of reductions in ambient PM2.5 concentrations upon just meeting the  current suite of
16    standards can impact the magnitude of additional risk reductions estimated for just meeting
17    alternative (lower) annual standard levels. Specifically, for those study areas with more peaky
18    PM2.s distributions43, application of peak shaving rollback will result in higher annual-average
19    PM2 5 levels remaining upon just meeting the current suite of standards.  If proportional rollback
20    is then used to simulate just meeting alternative annual standard levels, a greater degree of
21    reduction in annual-average PM2 5 concentrations will result, since the "starting point" for the
22    calculation (annual-average PM2 5 levels upon just meeting the current suite  of standards) will be
23    higher. This translates into larger reductions in risk simulated for alternative annual standards
24    when more localized patterns of PM2 5 reduction are involved with simulating just meeting the
25    current suite of standards. In instances where an urban study area has relatively peaky PM2.5
26    concentrations, the difference in projected risk reduction for an alternative annual standard can
27    be substantial. For example, with Los Angeles which has fairly peaky PM2 5 levels, we predict a
28    13% reduction in long-term exposure-related mortality with the 12 |ig/m3 alternative annual
29    standard (relative to risk under the current suite of standards) if we assume a regional pattern of
30    PM2 5 reductions for simulating just meeting the current standard as reflected in the application
      42 We note that additional variation in the risk estimates, in terms of both risk reduction across standard levels and
      residual risk for each of the alternative annual standard levels, results from differences across study areas in the
      relationship between the maximum monitor annual-averages values used in estimating percent reductions under an
      alternative standard and the composite monitor annual-average values used in estimating long-term exposure-related
      risk.
      43 The term "peaky" as used here, refers to urban study areas with relatively high 24-hour design values and lower
      annual average design values.
             Draft Do Not Quote or Cite           2-83                         March 2010

-------
 1    of proportional rollback. By contrast, we predict a 48% reduction in long-term exposure-related
 2    mortality risk if we assume a more localized pattern of reduction in ambient PM2.5 levels, as
 3    reflected in the application of peak-shaving rollback (US EPA 2010a, section 6.2.2).
 4          Regarding the magnitude of risk estimated to remain after simulation of alternative
 5    annual standards, the four study areas displaying the greatest degree of risk reduction across the
 6    range of alternative annual standard levels considered (Atlanta, Baltimore, Birmingham and
 7    Houston) have PM2.5-related IHD mortality estimates (under the lowest alternative annual
 8    standard of 12/35) ranging from 85-110 (Birmingham) to 220-280 (Houston) (US EPA, 2010a,
 9    Appendix E, Table E-21 and E-30). The two urban  study areas with the greatest degree of PM2.5-
10    related risk in absolute terms (Los Angeles and New York) do not exhibit significant reductions
11    in risk until the lowest annual standard level of 12/35  is considered, with PM2 5-related IHD
12    mortality estimated at 750-950 and 1,420-1,800, respectively under that alternative standard (US
13    EPA, 2010a, Appendix E, Table E-21 and E-30).  In terms of the percentage of long-term
14    mortality attributable to PM2.5, we see the following levels for the alternative annual standard
15    levels. For an annual standard level of 14 |ig/m3, the percent of total incidence of IHD mortality
16    attributable to PM2 5 in the 5 urban study areas experiencing risk reductions ranges from 9-11.3%
17    (Detroit) to 11.8-14.9% (Atlanta)  (US EPA 2010a, Appendix E, Tables E-24 and E-33). For an
18    annual standard of 12 |ig/m3, estimated risk remaining in the 12 urban study areas experiencing
19    risk reductions ranges from 6-7.6% (Phoenix) to 9-11.4% (Atlanta), again for PM2 5-attributable
20    long-term exposure-related mortality (US EPA 2010a, Appendix E, Tables E-24 and E-33).
21          We note that there is considerably less variability in the estimates of risk remaining after
22    simulation of the alternative annual standards levels, compared with the magnitude of risk
23    reductions across the study areas.  This reflects the fact that study areas experiencing risk
24    reductions under the alternative annual  standards will have similar annual-average PM2 5 levels
25    given that this simulation involves annual standard  levels that are controlling (i.e., these study
26    areas have their annual-average PM2.5 levels lowered to meet the same annual standard level).
27          Observations made earlier in section 2.1.1 in the context of the current suite of standards
28    regarding uncertainty and its impact on risk estimates apply in this context as well.  Specifically,
29    given the results of the sensitivity analysis examining the form of the C-R functions for long-
30    term exposure-related mortality, combined with only modeling risk down to the LML, we have
31    increased confidence that we have not overstated either the magnitude of risk reductions across
32    alternative standard levels, or the magnitude of risk remaining for a given standard level.
33    Risk reduction associated with alternative 24-hour standards (13/30 and 12/25)
34          We then focused on evaluating risks associated with long-term PM2 5 exposures when
35    varying the level  of the 24-hour standard while holding the level of the annual standard constant.

            Draft Do Not Quote or Cite           2-84                        March 2010

-------
 1    Comparing risks associated with just meeting the 13/35 and 13/30 suites of alternative standards
 2    (which reflect a 5 |ig/m3 reduction in the 24-hour standard, while holding the annual fixed at
 3    13|ig/m3), we see considerable variation in the magnitude of risk reduction across urban study
 4    areas. For example, St Louis, under the 13/35 suite of alternative standards has IHD mortality
 5    risk attributable to PM2.5 reduced by 22% relative to risk under the current suite of standards.
 6    Very little additional risk reduction (increasing from 22% to 24%) is estimated under the 13/30
 7    alternative suite of standards.  In contrast, with Salt Lake City, we estimate that the 13/35 suite of
 8    alternative standards will produce no risk reduction relative to the current suite of standards,
 9    while the 13/30 suite would produce a 55% reduction in IHD mortality risk relative to risk under
10    the current standard level (see Figure 2-3 and US EPA 2010a, Table E-27 in Appendix E).
11           The additional risk reduction provided by an alternative 24-hour standard is more
12    substantial in comparing the 12/25 and 12/35 alternative suites of standards,  although there is
13    also greater variability in the magnitude of risk reductions across study areas. For example,
14    Atlanta, which had a 32% reduction in risk under the 12/35  suite of standards (relative to the
15    current standard level) only sees an marginal increase to 34% under the 13/25 suite of standards.
16    By contrast,  Salt Lake City, which has no reduction in risk under the 12/35 suite of standards
17    (relative to the current suite of standards),  sees a -100% reduction in risk under the alternative
18    suite of 12/25 (see Figure 2-3 and US EPA, 2010a, Appendix E, Table E-27). The reduction for
19    Salt Lake City reflects a very high 24-hour design value which, when reduced to meet the 24-
20    hour standard of 25 |ig/m3  produced a substantial reduction in the annual design value (given
21    application of the proportional adjustment to simulate rollback), such that the value was very
22    close to 5.8 |ig/m3 (the LML below which long-term exposure-related mortality is not
23    estimated). The specific pattern of risk reduction reflects whether the 24-hour or annual standard
24    was controlling (see discussion below regarding patterns of risk reduction). We also note, that
25    under the 12/25 alternative suite of standards, we see that nine of the study areas (Detroit,
26    Fresno, Los Angeles, New  York, Philadelphia, Phoenix, Pittsburgh, Salt Lake City and Tacoma)
27    have reductions in risk that are at least twice as large as for the 12/35 suite of alternative
28    standards, with some substantially larger (see Figure 2-3 and US EPA, 2010a, Table E-27 in
29    Appendix E).
30           Regarding risk remaining after simulation of the suite of alternative 24-hour standards,
31    the four study areas displaying the greatest degree of reduction across these two alternative suites
32    of standards  (Tacoma, St. Louis, Los Angeles and Fresno), have PM2.s-related IHD mortality
33    estimates (under the 12/25 case) ranging from 3-4 (Tacoma) to 290-360 (Los Angeles) (US EPA,
34    2010a,  Appendix E, Table E-21 and E-30). The other urban study area with the greatest degree
35    of PM2.s-related risk in absolute terms besides New York (New York) has PM2.5-related all-cause
36    mortality estimated at 820-1,040 under the 12/25 case.  In terms of the percentage of long-term
             Draft Do Not Quote or Cite           2-85                         March 2010

-------
 1    mortality attributable to PM2.5, we see that under the 13/30 suite of standards, the percent of total
 2    incidence of IHD mortality attributable to PM2.5 ranges from 1.3 to 1.7%  (Salt Lake City ) to
 3    10.4 to 3.2% (Atlanta) (US EPA, 2010a, Appendix E, Tables E-24 and E-33). Under the 12/25
 4    alternative suite, risks for this metric range from ~0 to 0.3% (Salt Lake City) to 8.8 to 11.1%
 5    (Atlanta).
 6           The observations presented above highlight variability both in the magnitude of risk
 7    reduction as well as in the risk remaining upon simulation of just meeting alternative 24-hour
 8    standards. This reflects the fact that, as noted earlier, alternative 24-hour standards can produce
 9    different degrees of reduction in the annual-average PM2.5 concentrations, depending on the
10    relationship between 24-hour and annual design values at a particular location.  It is these annual -
11    average PM2 5 levels that drive changes in long-term exposure-related mortality. Those study
12    areas with relatively peaky PM2.5 levels, such as Salt Lake City, can see a substantial reduction in
13    annual-average PM2s levels under simulation of alternative 24-hour standards.  By contrast, study
14    areas with less peaky PM2.5 levels, such as St. Louis, see substantially smaller reductions in
15    annual-average PM2.5 levels under the suite of alternative 24-hour standards and consequently
16    1 ower degree s of ri sk reducti on.
17           In addition, as noted earlier, in those instances where alternative 24-hour standards
18    produce larger reduction in annual-average PM2 5 levels (and consequently greater reductions in
19    risk), often these reductions involve relatively low annual-average PM2 5 levels. In some cases,
20    such as with Salt Lake City, these annual-average PM2.5 levels can approach the LML, which
21    reduces our overall confidence in estimates of long-term exposure-related mortality for these
22    locations.
23           Observations made earlier regarding the impact of variability in simulating changes in
24    PM2.s distributions using different rollback approaches, and its impact on the degree of risk
25    reduction, also hold for these simulations of alternative 24-hour standards. Specifically, in those
26    instances where PM2.5 distributions are more peaky, application of peak shaving rollback
27    (reflecting more localized patterns of ambient PM2 5 reduction) would result in  smaller
28    reductions in annual-average PM2.5 concentrations and consequently, smaller reductions in
29    estimates of long-term exposure-related mortality. For example, with Salt Lake City, which has a
30    peaky PM2 5 distribution, under the 12/25  suite of standards application of proportional rollback
31    results in an annual average PM2.5 concentration of 5.7 |ig/m3, while application of peak shaving
32    results in an estimate of 8.9 |ig/m3 (as noted earlier, differences in annual-average PM2 5 levels
33    translate into differences in risk). In contrast, simulation of the 12/25 suite of standards for
34    Baltimore, which has a less peaky PM2.5 distribution, results in an annual average PM2.5
35    concentration of 10.7 |ig/m3 for proportional rollback compared to  10.8 |ig/m3 with peak

             Draft Do Not Quote or Cite           2-86                        March 2010

-------
 1    shaving, implying that there would be little difference in risk (see US EPA, 2010a, Table F-49 in
 2    Appendix F).
 3           Observations made earlier regarding overall confidence in the estimates of long-term
 4    exposure-related mortality also hold for these estimates (i.e., the sensitivity analysis results
 5    combined with the fact that we modeled risk down to LML result in our concluding that it is
 6    unlikely we have overstated either the degree of risk reduction or the degree of residual risk).
 7    Summary
 8           The results discussed above show that simulating just meeting alternative annual
 9    standard levels in the range of 14 to 12 |ig/m3 can produce substantial reductions in long-term
10    exposure-related mortality risk (with the magnitude of risk reduction increasing as lower annual
11    standards in this range are considered).  Furthermore, the results suggest that alternative 24-hour
12    standard levels in the range of 30 to 25 |ig/m3 can produce substantial reductions in estimated
13    risk, beyond that produced by simulations of just meeting lower annual standard level down to
14    12 |ig/m3 (combined with a 24-hour standard of 35 |ig/m3).  This results from the simulation of
15    the alternative 24-hour standard levels producing substantially lower annual-average PM2.5
16    concentrations, which drive reductions in both long-term and short-term exposure-related risk.
17    However, the results also show that there can be considerable variability across study areas in the
18    degree to which alternative 24-hour standard levels produce reductions in annual average PM2.5
19    concentrations and, consequently, reductions in risk. This variability depends on the peakiness
20    of the PM2.5 distribution in an area and on the pattern of reductions in ambient PM2.5 levels
21    associated with just meeting the current and alternative suites of standards. Conversely, the
22    analysis also suggests that more consistent annual-average PM2 5 concentrations, and thus more
23    consistent reductions in estimated risk and more uniform levels of  public health protection
24    would likely result from simulating just meeting alternative annual standards at levels below  12
25    |ig/m3 (i.e., below the lowest annual standard level considered in this assessment).
26           Furthermore, because the alternative 24-hour standard levels assessed resulted in lower
27    simulated annual-average PM2 5 levels (often approaching the LML used in modeling risk for
28    long-term exposure-related mortality), we also conclude that there is greater uncertainty
29    associated with risk estimates derived for the alternative 24-hour standards, relative to risks
30    estimated for the alternative annual standards.

31    2.3.5   Alternative Levels to Address Health Effects Related to  Short-term PM2.s Exposures
32           In considering alternative PM2.5  standards that would provide protection against health
33    effects related to short-term exposures, staff has taken into account both evidence-based and risk-
34    based considerations.  First, as discussed in section 2.1.3, it is our view that the evidence from
35    studies of short-term exposure to PM2.5  can help inform consideration of the protection afforded
             Draft Do Not Quote or Cite           2-87                         March 2010

-------
 1    by both a 24-hour and an annual standard (section 2.3.5.1).  Secondly, we have considered the
 2    results of the quantitative risk assessment presented in the second draft RA (US EPA, 2010a) to
 3    assess the extent to which alternative annual and/or 24-hour standards can be expected to reduce
 4    the estimated risks attributable to short-term exposure to PM2.5 (section 2.3.5.2).  With respect to
 5    considering the protection afforded by the suite of PM2.5 standards, our preliminary integrated
 6    conclusions presented in section 2.3.6 are based in part on the preliminary staff conclusions from
 7    this section and in part on preliminary staff conclusions from section 2.3.4 in which we discuss
 8    alternative standards to address health effects related to long-term PM2.5 exposures.

 9           2.3.5.1 Evidence-based Considerations
10           In taking into account evidence-based considerations for informing our understanding of
11    alternative levels to address health effects related to short-term fine particle exposures, we
12    initially focused on short-term multi-city PM2.5 exposure studies. We observe that multi-city
13    studies assessed PM2.5-associated health effects among larger  study populations, providing
14    enhanced power to detect PM2.5-associated health effects. In addition, short-term multi-city
15    PM2.5 exposure studies generally provide  spatial coverage for  different regions across the
16    country, reflecting differences in PM2.5 sources, composition, and potentially other factors which
17    might impact PM2.5-related risk. We have also evaluated short-term single-city PM2.5 exposure
18    studies that provide additional insights on associated health effects occurring at or below the
19    PM2.5 concentrations reported in the short-term multi-city PM2.5 exposure studies,  specifically in
20    areas that may have unusually high peak-to-mean ratios of PM2s concentrations, possibly
21    associated with strong local or seasonal sources of fine particles.
22           In considering a range of health outcomes, we place the greatest weight on associations
23    that have been judged in the ISA to be causal or likely causal, while being mindful of evidence
24    that informs our understanding of impacts on susceptible populations.44 We have  considered a
25    number of factors including: (1) the extent to which these studies report statistically significant
26    and relatively precise relative risk estimates; (2) the extent to which the reported associations are
27    robust to co-pollutant confounding; and (3) the extent to which the studies used relatively
28    reliable air quality data. In particular, we focused on those specific studies conducted in the U.S.
29    and Canada, briefly summarized in section 2.2.1 and presented in more detail in chapter 6 of the
30    ISA, that provide evidence of associations in areas that would likely have met the  current 24-
      44 In contrast to causal inferences for effects associated with long-term PM2 5 exposures, the ISA has not identified
      any health outcomes as having evidence suggestive of a causal association with short-term PM2 5 exposures. Thus,
      the discussion of evidence-based considerations in this section focuses on those effects for which the ISA has
      determined there is a causal or likely causal association with short-term PM2 5 exposures only.
             Draft Do Not Quote or Cite           2-88                         March 2010

-------
 1    hour PM2.5 standards during the time of the study.45 We believe that this body of evidence can
 2    serve as a basis for considering alternative 24-hour and annual PM2.5 standards that would
 3    provide appropriate protection for health effects associated with short-term PM2 5 exposures.
 4           Collectively, we have considered the evidence to address the following question:
 5    •  To what extent does the currently available evidence provide support for revising the
 6       current suite of standards to provide protection from short-term fine particle
 7       exposures?
 8           In considering what alternative levels would provide protection for effects associated
 9    with short-term PM2.5 exposure, we first recognize that the ISA has concluded that no discernable
10    threshold, or lowest-observed-effect levels, can be identified based upon the currently available
11    evidence (US EPA, 2009a. section 2.4.3). Being mindful  of the challenges posed by issues
12    relating to threshold and background concentrations, we have considered two alternative
13    approaches for using the evidence from short-term exposures studies to inform preliminary staff
14    conclusions regarding alternative standard levels. First, in considering an approach in which the
15    24-hour standard to set to be the primary means for providing protection for health effects
16    associated with short-term exposures to PM2.5, we considered evidence from short-term exposure
17    epidemiological studies as a basis for an alternative 24-hour PM2.5 standard. Specifically, we
18    focused on the upper end of the distributions of 24-hour PM2.5 concentrations, particularly in
19    terms of the highest 98th percentile values for a location, reflecting the form of the  current 24-
20    hour standard.
21            Alternatively, we have also considered an approach in which the annual standard is set to
22    be the primary means for providing protection for effects associate with both long  and short-term
23    PM2.s exposures.  This approach recognizes that, as noted in previous reviews, much of the risk
24    related to daily exposures, when aggregated on an annual basis, results from the large number of
25    days during which the 24-hour average concentrations are in the low- to mid-range of the entire
26    distribution. Thus, to reduce the aggregate short-term exposure-related risk, it is necessary to
27    shift the bulk of the distribution to lower concentrations, not just to limit the concentrations on
28    days when the PM2 5 concentrations are relatively high (US EPA, 2005, p. 4-67, Figure 4-10;
29    Abt, 1996, section 7.1, p.79, Exhibit 7.6). This approach reflects  consideration for allowing the
30    annual standard to serve in most areas as the target for control programs designed to be effective
31    in lowering the broad distribution of PM2.5 concentrations, thus protecting not only against
32    effects associated with long-term exposures but also effects associated with short-term
      45 As noted in section 2.3.4.1 above, we requested additional air quality distribution data from authors of specific
      epidemiological studies. To inform alternative standard levels to provide protection for effects associated with short-
      term PM2 5 exposures, we were particularly interested in 98th percentile and long-term mean PM2 5 concentrations
      considered in short-term exposure studies. In focusing on the current forms of the 24-hour and annual standards, we
      were also interested in information related to the highest reporting monitors.
             Draft Do Not Quote or Cite           2-89                         March 2010

-------
 1    exposures.  This judgment reflects the recognition that changes in PM2.5 air quality designed to
 2    meet an annual  standard would likely result not only in changes in lower annual average
 3    concentration but also in fewer and lower peak 24-hour concentrations.  In exploring this
 4    alternative approach, we considered the evidence from the currently available short-term
 5    exposure studies, specifically the long-term mean PM2.5 concentrations from these studies as a
 6    basis to inform preliminary conclusions for alternative levels for an annual PM2.5 standard.
 7    Furthermore, we recognize that if the primary purpose of the annual standard was to provide
 8    protection for effects associated with both long- and short-term exposures, the purpose of the 24-
 9    hour standard would than shift.  In that case, the primary purpose of the 24-hour standard would
10    be to serve to provide supplemental protection to provide additional protection beyond the
11    protection afforded by the annual standard in areas with unusually high peak-to-mean ratios of
12    PM2.5 concentrations, possibly associated with strong local or seasonal sources.
13           In first considering the approach of using the 24-hour standard as the primary means for
14    providing protection for health effects associated with short-term exposures to PM2.5, we look to
15    the 98th percentile air quality values for specific  short-term exposure studies to inform our
16    consideration of alternative standard levels (see Figure 2-4). With respect to mortality evidence
17    (causal association) from multi-city short-term exposure studies, we note that, as discussed in
18    section 2.2.1, Zanobetti and Schwartz (2009) and Burnett et al. (2004) reported a 24-hour 98th
19    percentile PM25 concentration averaged across all cities and all years of 34.3 |ig/m3 and 38.0
20    ng/m3, respectively. As presented in the ISA, despite large differences in air quality
21    concentrations,  Zanobetti and Schwartz (2009) reported "all-cause mortality risk estimates that
22    were fairly uniform across the climatic regions, except for the 'Mediterranean' region"46 (US
23    EPA, 2009a, p.  6-178, Figure 6-24).
24           With regard to the evidence from studies assessing short-term exposure to PM2 5-
25    associated with cardiovascular (causal association) and respiratory (likely causal association)
26    morbidity effects, recent multi-city analysis of MCAPS data by Dominici et al. (2006a) and Bell
27    et al. (2008) reported an overall 24-hour average 98th percentile PM2 5 concentration across all
28    cities and all years of 34.8 |ig/m3 and 34.2 |ig/m3, respectively.  As noted in section 2.2.1, an
29    unpublished sub-analysis of Dominici et al (2006a) restricted to days with 24-hour average
30    concentrations at or below the level of the current 24-hour standard (35  |ig/m3) indicated that, in
31    spite of a reduced statistical power from a smaller number of study days, statistically significant
32    associations were still observed between short-term PM2 5 exposure and hospital admissions for
33    cardiovascular and respiratory diseases (Dominici, 2006b).
      46 The Mediterranean region included cities in OR, CA, and WA.
             Draft Do Not Quote or Cite           2-90                         March 2010

-------
Study
Chimonas & Gessner (2007, 093261)
Lisabethetal. (2008, 155939)
Slaughter etal. (2005, 073854)
Rabinovitch et al. (2006, 088031)
Chen etal. (2004,087262)
Phon ot al tlftTfc fTfi7Q49'\
Funa etal. (2006,089789)
Villeneuve et al. (2003, 055051)
Stieb etal. (2000.011675)
Villeneuve et al. (2006, 090191)
Lin etal. (2005, 087828)
Mar etal. (2004.057309)
Rich etal. (2005, 079620)
Dockery et al. (2005, 078995)
Rabinovitch et al. (2004, 096753)
Pope etal. (2006, 091 246)
Slaughter et al. (2005, 073854)
Pope etal. (2008, 191969)
Zanobetti and Schwartz (2006, 090195)
Peters etal. (2001,016546)
Delfino etal. (1997, 082687)
Sullivan etal. (2005,050854)
Burnett etal. (2004,086247)
Bell etal. (2008. 156266)
Wilson etal. (2007, 157149)
Zanobetti & Schwartz (2009, 188462)
Burnett and Goldbera (2003, 042798)
Dominici et al. (2006, 088398)
Fairley (2003, 042850)
Zhanaetal. (2009, 191970)
O'Connor etal. (2008, 156818)
Klemm and Mason (2003, 042801)
Franklin etal. (2008,097426)
NYDOH (2006, 0901 32)
Ito etal. (2007, 156594)
Franklin etal. (2007,091257)
Rich etal. (2006,089814)
Svmons etal. (2006,091258)
Sheppard (2003, 042826)
NYDOH (2006, 0901 32)
Burnett etal. (1997.084194)
a uqlm'
Outcome
Asthma HA
LRIHA
IschemicStroke/TIAHA
Asthma Exacerbation
Asthma Medication Use
COPD HA
Respiratory HA
Nonaccidental Mortality
CVD ED Visits
Respiratory ED Visits
Hemhrgc Stroke HA
Ischemic Stroke HA
TIAHA
RTIHA
Respiratory Symptoms (any)
Respiratory Symptoms (any)
Ventricular Arrhythmia
Ventricular Arrhythmia
Asthma Exacerbation
IHDHA
CVD HA
Respiratory ED Visits
CHFHA
Ml HA
Pneumonia HA
Ml
Respiratory HA (summer)
Ml
Nonaccidental Mortality
Respir HA
CVD HA
CVD Mortality
Nonaccidental Mortality
Nonaccidental Mortality
CBVD HA
PVDHA
IHDHA
Dysrhythmia HA
CHFHA
COPD HA
RTIHA
Nonaccidental Mortality
ST Segment Depression
Wheeze/Cough
Nonaccidental Mortality
Nonaccidental mortality
Asthma ED Visits
Asthma HA
Non-accidental Mortality
Ventricular Arrhythmia
CHFHA
Asthma HA
Asthma ED Visits
Respiratory HA (summer)
CVD HA (summer)
Mean3
6.1
6.1
7.0°
7.3°
7.4
7.7
7.7
7.9
8.5
8.5
8.5
8.5
8.5
9.6
9.8"
9.8"
9.8°
10.3°
10.6"
10.8
10.8
10.8
11.1°
11.1°
12.1
12.1
12.8
12.8
12.9"
12.9d
13.0
13.2"
13.3
13.3
13.3
13.3
13.3
13.3
13.3
13.3
13.6
13.91
14.0C
14.7eJ
14.8
15.0k
15.1
15.6
16.2e
16.5"
16.7
16.71
16.8
98tha


23.6f

17.2r



27.3f
27.3f
24.0'
24.0'
24.0r

25.8'
25.8r
29.3f
29.6'
29.6'
44.5J


28.2'
31 .2r

38.0'
34.2'
34.2f
31 .6r
34.3'
38.9'
34.8'
34.8'
34.8'
34.8'
34.8r
34.8'
34.8'
59.0'
37.6'
39.0"

43.0f

39.0f
45.8r

50 1'
46.6'

47.4'
16.8 47.4'
Effect Estimate (95% Cl)
• i
, i
I i
I • 	
i»
r • 	
I,"
i




^_^_^__i__


	 - 	
t
— •!.


I s


If



— » 1 	 H









U«-
• -
9'
*
b
t ;
| .9.
1*
_ i
* J
•-#-
f |
	 1-* 	

h Averaged annual values for years in . 	 . 	 , 	 , 	 1 	 . 	 . 	 , 	 . 	 . 	 ,
b Study did not present mean; median presented.
c Mean estimated from data in study.
d Mean value slightly different from those reported in
  the published study or not reported in the published
  study; mean was either provided by study authors
  or calculated from data provided by study authors.
e Mean value not reported in study; median
  presented.
f 98th percentile of PM25 distribution was either
  provided by study authors or calculated from data
  provided by study authors.
g 98th estimated from data in study.
  study from data provided by study    '
  author.                       "'
i  Air quality data obtained from original
  study: Schwartz etal. (1996, 077325).
j  Mean PM25 concentration reported is
  for lag 0-2.
k  Bronx; TEOM data.
I  Manhattan; TEOM data.
                                                                                         0.8         1.0         1.2
                                                                                               Relative Risk / Odds  Ratio
1.4
1.6
                                               m Study does not present an overall effect
                                                 estimate; the vertical lines represent
                                                 the effect estimate for each of the areas
                                                 of Phoenix examined.
Figure 2-4.  Summary of Effect Estimates (per 10 jig/m3) and Air Quality Distributions for Short-term
Exposure Studies
Source:   US EPA, 2009a, Figure  2-1.

-------
 1           In looking at short-term single-city PM2.5 exposure studies conducted in areas that would
 2    have met the current level of the 24-hour standard, we recognize that these studies can provide
 3    additional insights regarding impacts on susceptible populations and/or on areas with isolated
 4    peak concentrations that could occur in areas that attain the current 24-hour and annual
 5    standards.  We first consider studies that report positive and statistically significant associations.
 6    One short-term single-city PM2.5 exposure study conducted since the last review reported a
 7    positive statistically significant association for short-term PM2.5 exposures in relation to
 8    respiratory symptoms among children in Phoenix with an average 98th percentile value of 25.8
 9    |ig/m3 (Mar et al., 2004). In addition, another short-term single-city PM2.5 exposure a study
10    noted in the last review also reported a positive statistically significant association between
11    short-term exposure to PM2 5 and myocardial infarction in Boston with an average 98th percentile
12    value of 28.2  |ig/m3 (Peters et al., 2001).
13           In also considering short-term single-city PM2 5 exposure studies that report positive but
14    statistically non-significant  associations for cardiovascular and respiratory endpoints in areas that
15    would have met the current 24-hour PM2.5 standard, we note a number of studies conducted in
16    Montreal (respiratory hospital admissions; Delfmo et al., 1997), Saint John (CVD  and respiratory
17    hospital admissions; Steib et al., 2000), Phoenix (CVD mortality; Wilson et al., 2007), Denver
18    (asthma medication use; Rabinovitch et al., 2006), Edmonton (hemorrhagic and ischemic stroke
19    hospital admissions; Villeneuve et al., 2006), and Nueces County, TX (ischemic stroke/transient
20    ischemic attack; Lisabeth et al., 2008). We note that the 98th percentile values reported in these
21    studies averaging across monitors in a city/county range from approximately 17.2  |ig/m3
22    (Rabinovitch  et al., 2006) in Denver to 31.6 |ig/m3 (Wilson et al, 2007) in Phoenix.  However,
23    other short-term single-city PM2.5 exposure studies report null findings for health effects noted
24    above (i.e., cardiovascular and respiratory morbidity) as being positively associated with short-
25    term exposure to PM2.5 in areas that would have met the current level of the 24-hour PM2.5
26    standard, including Phoenix (respiratory symptoms in adults; Mar et al., 2004); Spokane (CVD
27    hospital admissions and respiratory ED visits; Slaughter et al. 2005), Denver (asthma
28    exacerbation; Rabinovitch et al., 2004) and Edmonton (ischemic stroke and transient ischemic
29    attack hospital admissions; Villeneuve et al., 2006).
30           In considering information from both multi- and single-city short-term PM2 5 exposure
31    studies as summarized in Figure 2-4, we believe that the range of alternative 24-hour PM2.5
32    standards appropriate for consideration should extend below the range of 98th percentile values
33    averaged across cities from  the multi-city  studies identified above so as to provide protection
34    from effects associated with short-term PM2 5 exposure. In light of the mixed findings reported
35    in single-city  studies, particularly for studies conducted in areas such as Phoenix, Denver, and
36    Edmonton that report both positive and null findings, we place comparatively greater weight on
             Draft Do Not Quote or Cite           2-92                        March 2010

-------
 1    the results from multi-city studies for the reasons previously noted at the beginning of this
 2    section. Specifically, the heightened statistical power of multi-city studies to detect PM2.5
 3    associated health effects in aggregated analysis across cities and greater spatial coverage for
 4    different regions across the country to account for variations in PM2.5 composition, sources, and
 5    potentially other factors which might impact PM2.5-related risk. However, we note that 98th
 6    percentile values reported in single-city studies may provide additional insights for consideration
 7    in identifying alternative standards to provide protection for potentially susceptible populations.
 8    Specifically, as noted above, Mar et al. (2004) provides evidence of a statistically significant
 9    association between a 98th percentile PM2.5 level of 25.8 |ig/m3 and respiratory symptoms among
10    children.
11           In recognition that the current form of the 24-hour standard discussed in section 2.3.3.2
12    applies to the measurements made at each population-oriented monitor, OAQPS conducted
13    additional air quality analyses for selected multi-city short-term PM2.5 exposure studies to better
14    understand air quality distributions averaged across monitors by the study investigators in
15    comparison to air quality levels focused on the highest reporting monitors (Schmidt, 2010).
16    These analyses, as noted in section 23.4.1, focused on selected mortality and hospitalization
17    studies and, as presented in Table 2-5 and Figure 2-5, yielded composite monitor values close to
18    the 98th percentile concentrations reported to the EPA by the study authors (ranging from 34.2
19    |ig/m3 (Bell et al 2008) to 34.8 |ig/m3 (Dominici et al., 2006).47 As was expected, estimated air
20    quality levels for the highest reported monitors, ranging from 35.4 |ig/m3 (Bell et al, 2008) to
21    38.9 |ig/m3 (Dominici et al 2007) were higher than for the composite monitor values, ranging
22    from 33.5  |ig/m3 (Bell et al, 2008) to 36.5 |ig/m3 (Zanobetti and Schwartz, 2009). When more
23    study areas were considered and/or  more years of air quality data were evaluated, the difference
24    between the 98th percentile values for the composite monitor across cities compared to the 98th
25    percentile for the highest reporting monitor across cities was decreased.
26           Based on the epidemiological evidence and air  quality analyses summarized above,  in
27    considering the 24-hour standard to be the primary means to provide protection for short-term
28    PM2.5 exposures, we reach the preliminary conclusion that alternative levels could be considered
29    in the range of approximately 35 to  25 |ig/m3.  A level  selected at the upper end of this range
30    would reflect consideration of a level somewhat below the 98th percentile values for the highest
31    reporting monitors in the multi-city studies (ranging from about 35 to 39 |ig/m3) and just below
32    the 98th percentile values averaged across all cities and all years in these studies (in the range of
      47 As noted by Schmidt (2010), differences between 24-hour 98th percentile and long-term mean concentrations
      reported by the study authors and the parallel values in the EPA analyses may be related to the EPA assessment only
      encompassing Federal Reference and Federal Equivalent Method air quality data while the study authors may have
      included additional (non-reference/equivalent) data. Also, there may be additional differences in the sets of
      monitoring sites utilized due to uncertainties associated with the geographical area boundary definitions.
             Draft Do Not Quote or Cite           2-93                         March 2010

-------
1
2
Table 2-5.  Comparison of Air Quality Data for Selected Epidemiological Studies- 98 Percentile Values
                                                                                                                48
Author Reported Data
Study
MCAPS
-Original
-Extended
National
Mortality
Study
Cite
Dominici et al 2006
Bell et al 2008
Zanobetti &
Schwartz, 2009
Years
1999-2002
1999-2005
1999-2005
98%
34.8
34.2
34.3
Range
10-86
9-77
18-80
EPA Analysis (Schmidt 2010)
Composite Monitor
98%
34.5
33.5
35.6
Range
11-77
10-67
17-75
Maximum Monitor
98%
36.6
35.4
36.9
Range
13-92
12-79
19-79
    48 All concentrations reported as ug/m3.

           Draft Do Not Quote or Cite
                                                2-94
March 2010

-------
1
2
                                                                                                 th
Figure 2-5. Distribution of Study-specific 98  Percentile Concentrations






m
.S
"bb
a
1 *
1
o
a
o
O






1 0 0
9 5
9 0
8 5
8 0
7 5
7 0
6 5
6 0
5 5
5 0
4 5
4 0
3 5
3 0
2 5
2 0
1 5
1 0
5
0
                                                                                EPA
                                                                                 C M
                                                EPA     Study   EPA    EPA
                                                max    Author   CM     max
                                      |         Bell          ||Dominici06| |   Z  a  n  o  b  e tt i      |

                  Left plots represent the study author results; middle plots represent EPA replication results using a community monitor, CM, (spatial average) approach; and
                  right plots represent EPA replication results using a maximum value approach (max monitor 98th percentile for each year averaged over the study timeframe).
                  Black outlined boxes denote inter-quartile range and medians, stars denote means with cyan rectangles spanning +/- one standard deviation from the mean,
                  and dots represent the minimum and maximum values.
               Draft Do Not Quote or Cite
                                           2-95
March 2010

-------
 1    about 34 to 35 |ig/m3).  A level selected at the lower end of this range would reflect
 2    consideration of a level just below the lowest 98th percentile value associated with statistically
 3    significant effects in a single-city study reporting effects in children, identified as a susceptible
 4    population (Mar et al., 2004, 25.8 |ig/m3).
 5           Alternatively, in considering the approach of using the annual standard as the primary
 6    means for providing protection for both long- and short-term PM2.5 exposures, we look to the
 7    long-term mean PM2 5 concentrations from short-term PM2 5 exposure studies to inform
 8    considerations of alternative annual standard levels.  Using this approach, the same short-term
 9    multi- and single-city PM2.5 exposure studies briefly described above were evaluated.  In
10    focusing on multi-city short-term PM2.s exposure studies of cardiovascular and respiratory
11    mortality, the ISA concludes that associations are generally consistent and precise at long-term
12    mean PM2.5 concentrations of 12.8 |ig/m3 and above (see Figure 2-4; US EPA, 2009a, pp. 2-9 to
13    2-11). More specifically, with respect to mortality effects, Zanobetti and Schwartz (2009) report
14    associations with an overall annual average across all 112 cities of 13.2 |ig/m3  and Burnett et
15    al.(2004) report associations with an overall annual average across 12 Canadian cities of 12.8
16    ng/rn3. With regard to cardiovascular effects, Zhang et al. (2009) reported positive statistically
17    non-significant association between an average PM2.5 concentration of 13.9 |ig/m3 and ST
18    segment depression among participants of the Women's Health Initiative.
19           In considering PM2 5-related hospital  admissions for cardiovascular and respiratory
20    effects in older adults, an identified susceptible population, Dominici et al. (2006a) and Bell et
21    al. (2008) report an overall annual average across 204 and 202 counties of 13.3|ig/m3 and 12.9
22    |ig/m3, respectively (US EPA, 2009a, Figure 2-1).  In looking at short-term single-city PM25
23    exposure studies, similarly to the discussion of the 24-hour standard, Mar et al. (2004) (average
24    PM2.s concentration of 9.8 |ig/m3) and Peters et al. (2001) (average PM2.s concentration of 12.1
25    Hg/m3) report statistically significant associations for areas currently meeting the annual PM2.5
26    standard. Several short-term single-city PM2.5 exposure studies report positive but non-
27    statistically significant associations in areas that would have met the current level of the annual
28    standard for respiratory and cardiovascular morbidity and mortality endpoints associated with
29    long-term mean PM2 5 concentrations ranging from 7.0 |ig/m3 to 13.6 |ig/m3 (Figure 2-4).
30    However, as noted in the discussion of this evidence with regard to alternatives levels for the 24-
31    hour standard, investigators also reported null (Slaughter et al. 2005; Rabinovitch et al., 2004)
32    and mixed (Mar et al. 2004, Villeneuve et al., 2006) results for similar endpoints in which
33    positive associations were noted above for areas with annual  levels ranging from 8.5 |ig/m3 to
34    10.8 |ig/m3 (Figure 2-4). These inconsistent  findings suggest the potential for localized events
35    influencing the observed associations in single-city short-term PM2 5 exposure studies.
36    Consequently, we observe that the results of  short-term PM2.5 exposure studies could potentially
            Draft Do Not Quote or Cite           2-96                         March 2010

-------
 1    inform judgments on a 24-hour PM2.5 concentration set at the 98th percentile that provides
 2    adequate protection in areas with unusually high peak-to-mean ratios of PM2.5 levels. However,
 3    due to the possibility that these studies represent air quality distributions that may be influenced
 4    by localized events and not be representative of air quality across the country, we reach the
 5    preliminary conclusion that these single-city short-term PM2.5  exposure studies do not provide
 6    support for informing an alternative annual standard levels that would apply across the entire
 7    U.S.
 8          In considering not only the level but also the form of the annual standard, we
 9    acknowledge, as was discussed in section 2.3.4.1, that there  may be differences between mean
10    PM2.5 concentrations averaged across monitors within a city/county as is typically considered in
11    epidemiological studies,  compared to the current form of the annual PM2 5 standard which
12    typically focuses on  ambient measurements from the highest reporting community-oriented
13    monitor. As further  discussed in section 23.4.1, OAQPS conducted additional air quality
14    analyses to better understand air quality distributions  considered in selected short-term exposure
15    studies to compare long-term mean concentrations averaged across monitors in comparison to air
16    quality levels focused on the highest reporting monitors (Schmidt, 2010).  This analysis, as
17    presented in Figure 2-2, reported composite monitor values close to the long-term mean
18    concentrations reported by the study authors. As was  expected, estimated air quality levels for
19    the highest reported  monitors were higher, ranging from  13.4 |ig/m3 (Bell et al.., 2008) to!4.8
20    |ig/m3 (Dominici et al., 2007),  than for the composite  monitor values, ranging froml2.9 |ig/m3
21    (Bell et al.., 2008) to!3.7 |ig/m3 (Zanobetti and Schwartz, 2009). When more study areas were
22    considered and/or more years of air quality data were evaluated, the difference between the long-
23    term  mean concentrations for the composite monitor across cities compared to the long-term
24    mean concentrations for the highest reporting monitor across cities was decreased. Of particular
25    interest,  is that the difference between the composite monitor  and the high monitor is even less
26    when focusing on air quality estimates one standard deviation  below the long-term mean across
27    cities for each of these short-term exposure studies. With concentrations ranging from 10.2
28    |ig/m3 (Bell et al., 2008)  to 10.7 |ig/m3 (Zanobetti and Schwartz, 2009) based one standard
29    deviation below the long-term mean averaged across monitors and across cities/counties and
30    concentrations ranging from 10.3 |ig/m3 (Bell et al., 2008) to 11.0 |ig/m3 (Zanobetti and
31    Schwartz, 2009) based on one  standard deviation below the long-term mean at the highest
32    reporting monitor averaged across cities/counties.
33          In considering the currently available evidence, we reach the preliminary conclusion that
34    the short-term PM2.5 exposure  studies provide a basis  for considering alternative levels for the
35    annual standard within a range below 13 |ig/m3 down to about 10 |ig/m3.  A standard in the range
36    of 13 to  12 |ig/m3 would reflect placing greater weight on setting a level somewhat below the
            Draft Do Not Quote or Cite          2-97                         March 2010

-------
 1    long-term mean concentrations reported in the short-term exposure studies (ranging froml2.8
 2    ng/rn3 and above). A standard in the range of 11 to 10 |ig/m3 would be consistent with a
 3    judgment that, recognizing the serious nature of the effects and that no discernable threshold for
 4    these effects can be identified, appreciable weight should be accorded to considering the lower
 5    end of the interquartile range of long-term mean concentrations reported in these studies, or a
 6    range within one standard deviation around the study mean of either the composite  or the highest
 7    reporting monitor (ranging from about 10 to 11  |ig/m3).
 8          As recognized above, an annual standard used as the primary means for providing
 9    protection for effects associated with both long- and short-term PM2.5 exposures cannot be
10    expected to offer an  adequate margin of safety against the effects of all short-term PM2.5
11    exposures, especially in areas with high  peak-to-mean ratios of PM2 5 concentrations, possibly
12    associated with strong local or seasonal  sources of fine particles. In considering a basis for
13    setting an alternative standard levels for a 24-hour standard that would provide supplemental
14    protection against days with high peak concentrations associated with localized "hotspots" and
15    risk arising  from seasonal emissions that might not be well controlled by a national annual
16    standard, we intend to conduct air quality analyses of the relationships between annual and 24-
17    hour concentrations to include in the second draft PA.

18          2.3.5.2 Risk-based Considerations to Inform Alternative Levels
19          This discussion considers the impact of alternative suites of standard levels  on short-term
20    exposure-related mortality and morbidity. As such, it is intended to address the following policy-
21    related questions:
22    •  To what extent do alternative standards reduce estimated risks associated with short-
23       term PMi.s exposure compared to  risks associated with simulating air quality to just
24       meet the current standards?  What roles do the current 24 hour and annual standards
25       have in simulating the risks remaining upon just meeting the  alternative suites of
26       standards considered in the risk assessment? What level of confidence is associated with
27       these risk estimates?
28          As was done above for long-term exposure-related mortality,  we have divided the
29    discussion of risk for short-term exposure-related mortality and morbidity, between (a)
30    reductions in risk associated with simulation of the three alternative annual standard levels and
31    (b) reductions in risk associated with the two alternative 24-hour standard levels.
32    Risk reduction associated with alternative annual standards (14/35,13/35, and 12/35)
33          Reductions in risk for short-term exposure-related mortality and morbidity exhibits
34    considerable variability across study areas for the alternative suite of annual standards (compared
35    with risk under the current suite of standards) (see section 2.2.1.1). For the 12 urban study areas
            Draft Do Not Quote or Cite          2-98                       March 2010

-------
 1    experiencing reductions in short-term exposure-related CV mortality under the lowest alternative
 2    annual standard level assessed (12/35), reductions in risk compared to risk under the current suite
 3    of standards ranged from 5% (Phoenix) to 23% (Atlanta and Birmingham) (see Figure 2-3 and
 4    US EPA, 2010a, Appendix E, Table E-90). Estimates of risk reduction for CV-related HA's
 5    exhibited a similar pattern across the study areas for the lowest annual standard level assessed
 6    (see Figure 2-6 and US EPA, 2010a,  Appendix E, Table E-108).
 7          Regarding the magnitude of risk estimated to remain after simulation of the alternative
 8    annual standards, the four study areas displaying the greatest degree of reduction across the
 9    alternative annual standards (Atlanta, Baltimore, Houston and St. Louis) have PM2.5-related CV
10    mortality estimates (under the lowest alternative annual standard of 12/35) ranging from 25
11    (Atlanta) to 84 (St. Louis - see US EPA, 2010a, Appendix E, Table E-84). The urban study area
12    with the greatest degree of PM2.5-related risk in absolute terms (New York) did not exhibit
13    significant reductions in risk until the lowest annual standard level of 12/35 was considered, at
14    which point we estimate a reduction in PM2.5-related CV mortality of 424 cases (see US EPA,
15    2010a, Appendix E, Table E-84). In terms of the percentage of mortality attributable to short-
16    term PM2 5 exposure, the four study areas experiencing the greatest degree of risk reduction
17    across the four alternative annual standard levels had estimates (for the alternative annual
18    standard of 12 |ig/m3) ranging from 0.6% (Atlanta) to 1.3% (Baltimore - US EPA, 2010a,
19    Appendix E, Table E-87). For CV-related HA's related to short-term PM2 5 exposure, estimates
20    for this same group of four study areas (for the lowest alternative annual standard assessed)
21    ranged from 32 (Atlanta) to  142 (St. Louis - RA Appendix E, Table E-102). This translated into a
22    percentage of CV HA's attributable to PM2.5 ranging from 0.28% (Atlanta) to 1.07% (Baltimore
23    - US EPA, 201 Oa, Appendix E, Table E-105).
24          As with long-term exposure-related mortality, variation in risk reduction across urban
25    study areas for both short-term exposure-related mortality and morbidity reflect to a great extent
26    varying degrees of reduction in annual-average PM25 levels. This in turn results from the initial
27    simulation of the current 24-hour standard level which can produce varying degrees of reduction
28    in annual-average PM2.5 levels depending on a number of factors (i.e., peakiness of the PM2.5
29    monitoring data, specific mix of annual and 24-hour design values for a given location and the
30    spatial pattern of reduction in ambient PM2 5 levels that is assumed). We note however, that
31    there is greater variability in the levels  of risk remaining after simulation of alternative annual
32    standard levels for short-term exposure-related health endpoints compared with long-term
33    exposure-related mortality. This primarily reflects the fact that we are  able to specify C-R
34    functions separately by urban area or region for short-term exposure related mortality and
35    morbidity endpoints, while a single C-R function was used in modeling long-term exposure-
36    related mortality by  endpoint across all urban study areas.
            Draft Do Not Quote  or Cite           2-99                        March 2010

-------
1
2
                          Figure 2-6. Percent Reduction in Short-term Exposure-related Mortality and Morbidity Risk
                                                  (alternative standards relative to the current standards)
                (Note: inset shows PM2.5 related incidence and percent of total incidence for CV under the current suite of standards)
                            35%
                            30%
                          125%
                            20%
                        i £
                        w o
                        it
                          o
                          3
                        O c
                        £ 1 10%
                          0)
                          Q.
                                       Atlanta, GA 32 (-33-95); 0.8% (-0.8%-2.4%)
                                       Baltimore, MD 62 (-4-126); 1.6% (-0.1%-3.2%)
                                       Birmingham, AL  -1  (-42-40); 0% (-1.5%-1.5%)
                                       Dallas, TX 29 (-19-76); 0.8% (-0.5%-2.2%)
                                       Detroit, Ml 60 (-8-127); 1% (-0.1%-2.2%)
                                       Fresno, CA  12 (-9-33); 0.7% (-0.5%-2%)
                                       Houston,  TX 46  (-31-122); 0.9% (-0.6%-2.4%)
                                       Los Angeles, CA -30 (-132-72); -0.2% (-0.7%-0.4%)
                                       New York, NY 473  (276-668); 2.1% (1.2%-3%)
                                       Philadelphia, PA  84 (22-145); 2.1% (0.5%-3.6%)
                                       Phoenix, AZ 84  (-4-170);  1.3% (-0.1%-2.7%)
                                       Pittsburgh, PA 43 (-9-93);  1.1%  (-0.2%-2.3%)
                                       Salt Lake City, UT 9 (-2-20); 0.8% (-0.2%-1.7%)
                                       St. Louis, MO 106  (24-187); 1.9% (0.4%-3.3%)
                                       Tacoma, WA 11 (-6-27); 0.7% (-0.4%-1.8%)
                                       15/35*
                                                       14/35
13/35
12/35
13/30
12/25
 4
 5
 6
 1
 8
 9
10
                                                          Recent Air Quality, Current Standard and Alternative Standards
      *Based on Zanobetti and Schwartz (2009). The legend contains, for each urban area, the incidence estimate (and 95% CI) and the estimate of percent of total
      incidence (and 95% CI) under the current standards.
      **The current standards consist of an annual standard of 15 ug/m3 and a daily standard of 35 ug/m3. Combinations of an annual standard (n) and a daily standard
      (m) are denoted n/m in this figure.
               Draft Do Not Quote or Cite
                                                                            2-100
                 March 2010

-------
 1           As noted earlier, given increased emphasis placed in this analysis on long-term exposure-
 2    related mortality, the uncertainty analyses completed for this health endpoint category are
 3    somewhat more comprehensive than those conducted for short-term exposure-related mortality
 4    and morbidity, which to some extent reflects limitations in study data available for addressing
 5    uncertainty in the later category. However, C-R functions used in generating estimates for short-
 6    term exposure-related mortality and morbidity are well-supported in the literature.
 7    Risk reduction associated with alternative 24-hour standards (13/30 and 12/25)
 8           Comparing risks associated with just meeting the 13/35 and 13/30 suites of alternative
 9    standards (which reflect a 5 |ig/m3 reduction in the 24-hour standard, while holding the annual
10    fixed at 13 jig/m3), we see variation in the magnitude of risk reduction across urban study areas.
11    For example, Baltimore, CV mortality risk attributable to PM2 5 under the 13/35 suite of
12    alternative standards was reduced by 14% relative to risk under the current suite of standards.
13    Very little additional risk reduction (increasing from 14% to 15%) is estimated under the 13/30
14    alternative suite of standards.  In contrast, with Salt Lake City, we estimate that the  13/35 suite of
15    alternative standards will produce no risk reduction relative to the current suite of standards,
16    while the 13/30 suite would produce a 15% reduction in CV mortality risk relative to risk under
17    the current standard level (see Figure 2-6 and US EPA, 2010a, Table E-90 in Appendix E).
18           The additional risk reduction provided by an alternative 24-hour standard is  more
19    substantial in comparing the 12/25 and 12/35 alternative suites of standards, although there is
20    also greater variability in the magnitude of risk reductions across study  areas. For example,
21    Atlanta, which had a 23% reduction in risk under the 12/35  suite of standards (relative to the
22    current standard level) only sees a marginal increase to 24% under the 13/25 suite of standards.
23    By contrast, Salt Lake City, which has no reduction in risk under the 12/35 suite of  standards
24    (relative to the current suite of standards), sees a 29% reduction in risk under the  alternative suite
25    of 12/25 (see Figure 2-3 and US EPA, 2010a, Appendix E, Table E-90). The pattern of risk
26    reduction for CV-related HA's is very similar to what is presented here for CV-related mortality
27    (see Figure 2-6 and US EPA, 2010a, Appendix E, Table E-108).
28           We note that while considerable reductions in risk are seen for short-term exposure-
29    related mortality across many of the study areas, these are smaller than reductions seen for long-
30    term exposure-related mortality. This reflects primarily the fact that risk is modeled down to
31    PRB for short-term exposure-related mortality, while it is only modeled down to LML for long-
32    term exposure-related mortality. Therefore, the incremental reductions in risk associated with
33    short-term exposure-related mortality when we consider alternative suites of standards involve a
34    smaller fraction of total risk compared with long-term exposure-related mortality.
            Draft Do Not Quote or Cite          2-101                        March 2010

-------
 1          Regarding risk remaining after simulation of the suite of alternative 24-hour standards,
 2    PM2.s-related CV mortality estimates (under the 12/25 case) ranging from 6 (Salt Lake City) to
 3    336 (New York) (see US EPA, 2010a, Appendix E, Table E-84). Estimates of CV-related HA
 4    under the 12/25 suite of standards range from 7 (Salt Lake City) to 534 (New York) (see
 5    Appendix E, Table E-102). In terms of the percentage of short-term mortality attributable to
 6    PM2.5, we see that under the 13/30 suite of standards, the percent of total incidence of CV
 7    mortality attributable to PM2.s ranges from -0% (Los Angeles, Birmingham) to 1.8% (New
 8    York, Philadelphia) (US EPA, 2010a, Appendix E, Table E-87). Under the 12/25 alternative
 9    suite, risks for this metric range from -0% (Birmingham, Los Angeles) to 1.1% (Baltimore).
10          The observations presented above highlight variability both in the magnitude of risk
11    reduction as well as in the risk remaining upon simulation of just meeting alternative 24-hour
12    standards. This reflects the fact that, as noted earlier,  alternative 24-hour standards can produce
13    different degrees of reduction in the annual-average PM2.5 concentrations, depending on the
14    relationship between 24-hour and annual design values at a particular location. As noted before,
15    it is these annual-average PM2.5 levels that drive changes in long-term exposure-related
16    mortality. In addition, as noted earlier, modeling of risk for short-term exposure-related mortality
17    and morbidity involves a combination of urban study area- and regional-specific C-R functions,
18    which adds additional variability to risk estimates generated across the study areas.
19          As noted earlier, given increased emphasis placed in this analysis on long-term exposure-
20    related mortality, the uncertainty analyses completed for this health endpoint category are
21    somewhat more comprehensive than those conducted for short-term exposure-related mortality
22    and morbidity, which to some extent reflects limitations in study data available for addressing
23    uncertainty in the later category. However, C-R functions used in generating estimates for short-
24    term exposure-related mortality and morbidity are well-supported in the literature.
25    Summary
26           The results discussed above show that simulating just meeting alternative annual
27    standard levels in the range of 14 to 12 |ig/m3 can produce substantial reductions in short-term
28    exposure-related mortality and morbidity risk (with the magnitude of risk reduction increasing as
29    lower annual standards in this range are considered). Furthermore, the results suggest that
30    alternative 24-hour standard levels in the range of 30 to  25 |ig/m3 can produce additional
31    reductions in estimated risk, beyond that produced by simulations of just meeting lower annual
32    standard level down to 12 |ig/m3 (combined with a 24-hour standard of 35  |ig/m3). However, we
33    note that the magnitude of estimated reductions in short-term exposure-related mortality are
34    substantially lower than estimates for long-term exposure-related mortality (when comparing
35    risks for the same urban study area and suite of standards).

            Draft Do Not Quote or Cite          2-102                        March 2010

-------
 1           In addition, while we would expect alternative 24-hour standards to produce more
 2    variable reduction in risk estimates and consequently in public health protection relative to
 3    alternative annual standards (for the same reasons presented earlier for long-term exposure-
 4    related mortality), we see that risk estimates for both annual and 24-hour standards are fairly
 5    variable for short-term exposure-related health effects categories. This reflects in part the fact
 6    that we use urban area-differentiated and regionally-differentiated C-R functions in modeling
 7    short-term exposure-related mortality and morbidity, which introduces additional variability
 8    across study areas in terms of both risk reduction and the magnitude of risks remaining after
 9    simulation of alternative suites of standards.

10    2.3.6   Preliminary Staff Conclusions on Alternative Levels to Address Health Effects
11           Related to Long- and Short-term PM2.s Exposures
12           In reaching preliminary staff conclusions on alternative standard levels to address health
13    effects related to both long- and short-term PM2.5 exposures, we have considered the currently
14    available scientific information including: epidemiologic evidence, including evidence of effects
15    in susceptible populations; air quality analyses; and estimates of risk reductions associated with
16    alternative annual and/or 24-hour standard levels, as well as the related limitations and
17    uncertainties associated with this information as presented in sections  2.3.4 and 2.3.5 and
18    discussed more fully in the ISA and second draft RA (US EPA, 2009a; US EPA, 2010a). As
19    outlined in section 2.1.3, we believe it is most appropriate to consider  the protection against
20    PM2.5-related mortality and morbidity effects,  associated  with long- and short-term exposures,
21    afforded by the annual and 24-hour standards taken together, rather than to consider each
22    standard separately. The extent to which these two standards are interrelated in any given area
23    depend in large part on the relative levels of the standards, the peak-to-mean ratios that
24    characterize air quality patterns in an area, and whether changes in air quality  designed to meet a
25    given suite of standards is of a more regional or localized nature.
26           In looking first at the nature of the associations between long-  and short-term PM2.5
27    exposures and a range of health outcomes, we have considered not only the causal inferences
28    presented in the ISA but also at what air quality concentrations these effects have been observed.
29    In evaluating these data, we are mindful that the ISA has  concluded that no discernable threshold
30    for an effect associated with long- or short-term PM2 5 exposures can be identified based on the
31    currently available evidence (US EPA, 2009a, section 2.4.3). We have considered a number of
32    different air quality metrics to inform our preliminary conclusions regarding alternative levels
33    that should be considered including:
34         •  Long-term mean ambient PM2 5 concentrations reported in long- and short-term
35            exposure studies;

             Draft Do Not Quote or Cite           2-103                       March 2010

-------
 1         •  Ambient PM2.5 concentrations down to one standard deviation below the long-term
 2            mean concentrations or the lower end of the interquartile range of air quality data
 3            evaluated in the long- and short-term exposure studies;

 4         •  The range of long-term PM2.5 mean concentrations and the point were the data become
 5            more sparse or where the confidence interval becomes notably wider, suggestive of a
 6            concentration below which the association becomes appreciably more uncertain and the
 7            possibility that an effects threshold may exist becomes more likely;

 8         •  Differences between long-term mean PM2 5 concentrations averaged across monitors
 9            within a city/county as is typically considered in epidemiologic  studies, compared to
10            the current form of the annual PM2.5 standard which typically focuses on ambient
11            measurements from the highest reporting  community-oriented monitor;

12         •  98th percentile 24-hour ambient PM2.5 concentrations identified for short-term exposure
13            studies; and

14         •  Differences between 98th percentile 24-hour ambient PM2 5 concentrations averaged
15            across monitors within a city/county compared to the current form of the 24-hour PM2.5
16            standard which focuses on ambient measurements from the highest reporting
17            community-oriented monitor
18          In considering the results of quantitative risk assessments conducted for this and previous
19    reviews, specifically the estimated risk remaining upon simulation of just meeting alternative
20    suites of standards,  we note the following:
21          •   Long-term exposure to PM2.5 has been shown to produce substantially larger mortality
22              risk (in terms of overall incidence and percent of total mortality) compared to short-
23              term PM2.5 exposure (section 2.2.2; US EPA, 2010a, p. 6-1).
24
25          •   Much of the risk related to daily exposures, when  aggregated on an annual basis,
26              results from the large number of days during which the 24-hour average
27              concentrations are in the low- to mid-range of the  entire distribution.  Thus, to reduce
28              the aggregate short-term exposure-related risk, it is necessary to shift the bulk of the
29              distribution to lower concentrations, not just to limit the concentrations on days when
30              the PM2.5 concentrations are relatively high (US EPA, 2005, p. 4-67, Figure 4-10;
31              Abt, 1996, section 7.1, p.79, Exhibit 7.6).
32
33          •   Alternative annual standard levels produced more consistent levels of risk reduction
34              (and consequently public health protection) with generally  higher levels of confidence
35              when compared with alternative 24-hour standards. Considerable variability exists
36              across urban study areas with respect to  the degree to which alternative 24-hour
37              standard levels produce reductions in annual average PM2 5 concentrations and,
38              consequently, reductions in risk (section 2.3.4.2).
39
40          Taken together, we believe the best way to provide requisite protection for effects
41    associated with both long- and short-term PM2.5 exposures is to define a suite of standards that
42    will provide generally consistent protection across the country.  In considering the roles that each

            Draft Do Not Quote or Cite           2-104                       March 2010

-------
 1    standard might be expected to play in the protection afforded by alternative suites of standards,
 2    we believe it is appropriate to select a policy approach where the annual standard is the
 3    "generally controlling" standard. This approach would reflect consideration for allowing the
 4    annual standard to serve in most areas as the target for control programs designed to be effective
 5    in lowering the broad distribution of PM2.5 concentrations, thus protecting not only against
 6    effects associated with long-term exposures but also effects associated with short-term
 7    exposures.  This approach reflects the recognition that changes in PM2.5 air quality designed to
 8    meet an annual standard would likely result not only in changes in lower annual average
 9    concentrations but also in fewer and lower peak 24-hour concentrations.
10           In addition, we recognize that an annual standard cannot be expected to offer an adequate
11    margin of safety against the effects of all  short-term PM2 5 exposures, especially in areas with
12    unusually high peak-to-mean ratios of PM2.5 levels, possibly associated with strong local or
13    seasonal sources, or for potential PM2.5-related effects that may be associated with shorter-than-
14    daily exposure periods (noted above in section 2.3.2).  As a result, in conjunction with an annual
15    standard that may be adopted as the generally controlling standard, in part to provide protection
16    against effects associated with short-term exposures, we  believe it is appropriate to use the 24-
17    hour PM2.5 standard to provide supplemental protection against days with high peak
18    concentrations associated with localized "hotspots" and risk arising from seasonal emissions that
19    might not be well controlled by  a national annual standard.
20           Thus, we reach the preliminary conclusion that it is appropriate to consider defining the
21    annual PM2 5 standard as the generally controlling standard and the 24-hour PM2 5 standard as a
22    "backstop" to provide additional protection, where needed.  We believe selecting a suite of
23    standards using this approach will provide a more uniform level of protection across the U.S.
24    compared to using an approach where the 24-hour standard  is the generally controlling standard.
25    This preliminary conclusion is consistent with the approach used to set the original PM2.5
26    standards in 1997 (see section 2.1.1.1).
27           In integrating the preliminary staff conclusions reached in sections 2.3.4 and 2.3.5
28    regarding alternative levels to provide protection for long- and short-term PM2.5 exposures,
29    respectively, with the proposed approach outlined above, we have reached  preliminary
30    conclusions  regarding alternative levels for both the annual  and 24-hour standards.  The body of
31    scientific evidence upon which these preliminary conclusions are based includes consideration of
32    effects in susceptible populations, including evidence of mortality (causal association),
33    cardiovascular effects (causal association) and/or respiratory effects  (likely causal association) in
34    various populations including the life stages of children and older adults, people with pre-
35    existing cardiovascular and respiratory disease, and people with lower socioeconomic status. As
36    outlined in section 2.3.4.1, in considering alternative levels  for an annual standard that would
             Draft Do Not Quote or Cite           2-105                        March 2010

-------
 1    provide protection with an adequate margin of safety, we believe it is also appropriate to take
 2    into account evidence of effects for which the ISA has determined the evidence is suggestive of
 3    an association with long-term PM2 5 exposure including reproductive and developmental effects
 4    and carcinogencity, mutagenicity, and genotoxicity.  At this time, PM2.5 concentrations reported
 5    in these studies are equal to or greater than ambient concentrations identified in studies reporting
 6    associations with mortality, cardiovascular effects, or respiratory effects.  Therefore, in selecting
 7    alternative levels, we note that, in providing protection for mortality and cardiovascular and
 8    respiratory effects, it is reasonable to  anticipate that protection will also be provided for a
 9    broader range of health outcomes.
10           Based on the currently available scientific evidence, we believe there is support for a
11    "generally controlling" PM2 5 annual standard in the range of 13 to 10  |ig/m3 to provide
12    protection for effects associated with  both long- and short-term PM2.5 exposures in conjunction
13    with a generally "non-controlling" 24-hour PM2 5 standard in the range of approximately 30 to 35
14    |ig/m3 to limit peak concentrations in areas with relatively high peak-to-mean PM2.5 ratios.
15    Selecting a suite of standards with an annual  standard in the range of 13 to 12 |ig/m3 in
16    conjunction with a 24-hour standard in the range of 35 to 30 |ig/m3 would reflect placing greater
17    weight on setting standard levels just  below the PM2.5 concentrations (long-term mean and 98th
18    percentile value) reported in the epidemiologic  studies reporting associations with mortality and
19    cardiovascular and respiratory effects. It would also recognize consideration for setting a
20    standard with an adequate margin of safety by selecting standard levels based on ambient
21    concentrations averaged across monitors as is typically done in epidemiologic studies and
22    applying those levels using the forms of the current annual and 24-hour PM2 5 standards which
23    generally focus on the ambient concentrations reported at the highest reporting monitors.
24           Alternatively, selecting a suite of standards with an annual standard in the range of 11 to
25    10 |ig/m3 in conjunction with a 24-hour standard of 30 to 25 |ig/m3 would reflect a more
26    precautionary approach. This approach would be consistent with a judgment that, recognizing
27    the serious nature of the effects and that no discernable threshold for these effects can be
28    identified, appreciable weight should  be accorded to recognizing that health effects may  occur
29    over the full range of concentrations observed in the epidemiologic studies. Using this approach,
30    we believe that it is reasonable to conclude that the evidence of association is strongest down to
31    somewhat further below the aggregate long-term mean concentrations reported in the long- and
32    short-term exposure studies, such  as down to one standard deviation below the  mean or to the
33    lower end of the interquartile range, which includes the range in which the data in the study are
34    most concentrated. To serve as an effective "backstop" for an alternative annual standard set at
35    the lower end of the preliminary range, we reach the preliminary conclusion that it would be

             Draft Do Not Quote or Cite          2-106                       March 2010

-------
 1    appropriate to set the level of the 24-hour PM2.5 standard somewhat lower than the current
 2    standard.
 3          We recognize that it is the link between the form and the level of the standard that
 4    determines the degree of public health protection the standard affords. As discussed in section
 5    2.3.3.1, the current form of the annual standard allows for spatial averaging across monitors
 6    within constraints that were narrowed in 2006. In light of the  currently available scientific
 7    evidence discussed in section 2.2.1 providing stronger support that socioeconomic status (SES)49
 8    modifies the association between PM and morbidity and morality outcomes, we believe that
 9    eliminating the spatial averaging provisions from the form of the annual PM2.5 standard in
10    conjunction with consideration of alternative levels for the annual standard discussed above is
11    appropriate for identifying a standard that provides requisite protection with an adequate margin
12    of safety. In the second draft of this Policy Assessment, we intend to explore more fully
13    information on the peak to mean air quality concentrations relating the 24-hour 98th percentile
14    PM2.5 concentrations relative to the mean annual concentrations to better inform our
15    understanding of the implications of retaining or modifying  the current form of the annual
16    standard.

17    2.4    PRELIMINARY STAFF CONCLUSIONS ON ALTERNATIVE FINE PARTICLE
18          STANDARDS
19          In reaching preliminary conclusions on potential alternative standards to provide requisite
20    protection for health effects associated with long- and  short-term fine particle exposures, staff
21    has considered these standards in terms of the basic elements of the NAAQS: indicator,
22    averaging time, form, and level.  In considering the scientific and technical information, we
23    reflect upon the information available in the last review integrated with information that is newly
24    available since the last review as assessed and presented in the ISA and second draft RA (US
25    EPA, 2009a; US EPA 2010a) and as summarized in sections 2.2 and 2.3.
26          As outlined in sections 2.1.3 and 2.3.6, in this review, our approach to reaching
27    conclusions about the adequacy  of the current suite of PM2.5 standards and potential alternative
28    standards that are appropriate for consideration is broader and more integrative than approaches
29    used in past reviews. In applying this approach, we recognize that there are various ways to
30    combine the suite of standards to achieve an appropriate degree of public health protection.  Such
31    an approach to standard setting, which integrates a much expanded body of health effects
32    evidence, more extensive air quality data and analyses, and a more comprehensive quantitative
33    risk assessment and  considers the combined protection afforded by the annual and 24-hour
34    standards,  has the potential to result in a more effective and  efficient suite of standards than an
      49 Measured using surrogates such as educational attainment, residential location, and income level.
            Draft Do Not Quote or Cite          2-107                        March 2010

-------
 1    approach that only considers long- and short-term exposures evidence, analyses, and standards
 2    independently.
 3           In presenting our preliminary conclusions regarding alternative suites of primary
 4    standards and ranges of levels for consideration, we summarize conclusions presented in sections
 5    2.3.1 through 2.3.6. We emphasize that these are preliminary conclusions that reflect
 6    consideration of the scientific  and technical information assessed and presented in the ISA and
 7    second draft RA (US EPA, 2009a; US EPA, 201 Oa).  We note that staff conclusions to be
 8    presented in the final RA and second draft PA will be based, in part, on input received from
 9    CASAC and the public on the second draft RA and this first draft PA, as well as additional
10    analyses that we are conducting that will help inform our consideration of alternative forms and
11    levels as noted in sections 2.3.3, 2.3.4, and 2.3.5.
12           We recognize that selecting from among alternative standards will necessarily reflect
13    consideration of the qualitative and quantitative uncertainties inherent in the relevant evidence
14    and in the assumptions that underlie the quantitative risk assessment. In identifying these
15    alternative suites of primary standards and ranges of levels for consideration, we are mindful that
16    the Clean Air Act requires standards to be set that are requisite to protect public health with an
17    adequate margin of safety, such that the standards are to be neither more nor less stringent than
18    necessary. Thus, the Act does not require that the NAAQS be set at zero-risk levels, but rather at
19    levels that avoid unacceptable risks to public health.
20
21       ( 1)Consideration should be given to revising the current PM2.5 primary standards to provide
22           increased public health protection from the effects of both long- and short-term exposures
23           to fine particles in the ambient air. This preliminary conclusion is based in general on the
24           evaluation in the ISA of the currently available epidemiologic, toxicologic, dosimetric,
25           and exposure-related evidence, and more specifically on the evidence of mortality and
26           cardiovascular  and respiratory morbidity effects in  areas where the current standards
27           were met, together with judgments as to the public  health significance of the estimated
28           incidence of effects upon just meeting the current suite of standards.
29
30       ( 2)The indicator for fine particle standards should continue to be PM2.5. This observation
31           reflects our preliminary conclusions that the available evidence does not provide a
32           sufficient basis for replacing or supplementing the PM2 5 indicator with an indicator
33           defined in terms of ultrafine particles or for any specific fine particle component or
34           source category of fine particles, nor does it provide a basis for excluding any component
35           or source category from the mix of particles included in the PM2.5 indicator.
36
            Draft Do Not Quote or Cite          2-108                        March 2010

-------
 1       ( 3) Averaging times for PM2.5 standards should continue to include annual and 24-hour
 2          averages to protect against health effects associated with long-term (seasons to years) and
 3          short-term (hours to days) exposure periods.  Consideration of other averaging times,
 4          including an averaging time less than 24 hours to address health effects associated with
 5          subdaily fine particle exposures or a longer averaging time to address effects associated
 6          with seasonal fine particle exposures, was limited by the relatively small amount of
 7          relevant information available in this review.
 8
 9       ( 4) Consideration should be given to retaining or revising the form of the annual standard.
10          Consideration should be given to revising the form of the annual standard to one based on
11          the highest community-oriented monitor in an area rather than a form that would allow
12          averaging across monitors (e.g., spatial averaging). The form of the 24-hour standard
13          should continue to be defined in terms of the 98th percentile of the distribution of 24-hour
14          PM2.5 concentrations, averaged over three years.
15
16       ( 5) Consideration should be given to alternative suites of PM2 5 standards to provide
17          protection against effects associated with both long- and short-term exposures, taking into
18          account both evidence-based and risk-based considerations. Further, we conclude it is
19          appropriate to consider setting the levels of the annual and 24-hour standards such that
20          the annual standard would be the "generally controlling" standard to provide protection
21          for both long- and short-term PM2 5  exposures in conjunction with a 24-hour standard to
22          provide supplemental protection against days with high peak concentrations associated
23          with localized "hotspots" and risk arising from seasonal emissions that might not be well
24          controlled by a national annual standard.  Integrated preliminary  conclusions on ranges
25          of alternative suites of standards that, when considered together, would provide requisite
26          protection against effects associated with both long- and short-term exposures include:
27
28          (a) Consideration of a revised annual PM2.5 standard within the range of 13 to 12 |ig/m3,
29              together with either retaining or revising the 24-hour PM2 5 standard within the range
30              of35to30|ig/m3.
31          (b) Consideration of a revised annual PM2 5 standard, within the range of 11 to 10  |ig/m3,
32              together with revising the 24-hour PM2.s standard within a range of 30 to 25 |ig/m
33
3
34    2.5    KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH AND DATA
35          COLLECTION
36          [This topic will be discussed in the second draft Policy Assessment.]
            Draft Do Not Quote or Cite          2-109                        March 2010

-------
  1    2.6    REFERENCES

  2    Abt (1996).  A Paniculate Matter Risk Assessment for Philadelphia and Los Angeles.  Abt Associates. Bethesda,
  3            MD. Update of July 3, 1996 (November 1996). Supplement to November 1996 Report (December 1996).
  4            Prepared for the Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
  5            Contract No 68-W4-0029. Available: http://www.epa.gOv/ttn/naaqs/standards/pm/s jm_prtd.html.

  6    Abt (2005).  Paniculate Matter Health Risk Assessment for Selected Urban Areas.  Abt Associates. Final Report.
  7            Bethesda, MD. Prepared for the Office of Air Quality Planning and Standards, U.S. Environmental
  8            Protection Agency, Contract No. 68-D-03-002. EPA 452/R-05-007A. Available:
  9            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm  cr td.html.

10    ATSDR (2006). A study of ambient air contaminants and asthma in New York City: Part A and B. Agency for
11            Toxic Substances and Disease Registry; Public Health Service; U.S. Department of Health and Human
12            Services. Atlanta, GA.
13            http://permanent.access.gpo.gov/lps88357/ASTHMA  BRONX FINAL REPORT.pdf.

14    Bell ML, Ebisu K, Peng RD, Walker J, Samet JM, Zeger SL, Dominic F. (2008). Seasonal and regional short-term
15            effects of fine particles on hospital admissions in 202 U.S. counties, 1999-2005. Am J Epidemiol,  168:
16            1301-1310.

17    Bell ML, Ebisu K, Peng R, Samet J, Dominici F. (2009). Hospital Admissions and Chemical Composition of Fine
18            Particle  Air Pollution. Am J Respir Crit Care Med, 179: 1115-1120.

19    Beelen R, Hoek G, van den Brandt PA, Goldbohm RA, Fischer P, Schouten LJ, Jerrett M, Hughes E, Armstrong B,
20            Brunekreef B (2008). Long-term effects of traffic-related air pollution on mortality in a Dutch cohort
21            (NLCS-AIR study). Environ Health Perspect, 116: 196-202.

22    Burnett RT,  Goldberg MS. (2003) Size-fractionated paniculate mass and daily mortality in eight Canadian cities. In:
23            Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:  Health
24            Effects Institute; pp. 85-90. Available: http://www.healtheffects.org/news.htm [16 May, 2003].

25    Burnett RT,  Stieb D, Brook JR, Cakmak S, Dales R, Raizenne M, Vincent R, Dann T. (2004). Associations between
26            short-term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup Health, 59:
27            228-236.

28    Chen Y, Yang Q, Krewski D, Shi Y, Burnett RT, McGrail K (2004). Influence of relatively low level of paniculate
29            air pollution on hospitalization for COPD in elderly people. Inhal Toxicol, 16: 21-25.

30    Chen LH, Knutsen SF, Shavlik D, Beeson WL, Petersen F, Ghamsary M, Abbey D. (2005). The association between
31            fatal coronary heart disease and ambient paniculate air pollution: Are females at greater risk? Environ
32            Health Perspect, 113: 1723-1729.

33    Chuang KJ,  Chan CC, Shiao GM,  Su TC (2005). Associations between submicrometer particles exposures and
34            blood pressure and heart rate in patients with lung function impairments. J Occup Environ Med, 47: 1093-
35            1098.

36    Dales R, Liu L, Szyszkowicz M, Dalipaj M, Willey J, Kulka R, Ruddy TD (2007). Paniculate air pollution and
37            vascular reactivity: the bus stop study. Int Arch Occup Environ Health, 81: 159-164.

3 8    Delfino RJ, Murphy-Moulton AM, Burnett RT, Brook JR, Becklake MR (1997). Effects of air pollution on
39            emergency room visits for respiratory illnesses in Montreal, Quebec. Am J Respir Crit Care Med,  155: 568-
40            576.

41    Dockery DW, Pope CA III, Xu X, Spengler JD, Ware JH, Fay ME, Ferris BG Jr, Speizer FE. (1993). An association
42            between air pollution and mortality in six US cities. NEJM 329: 1753-1759.

              Draft Do Not Quote or Cite           2-110                          March 2010

-------
  1    Dockery DW, Cunningham J, Damokosh AI, Neas LM, Spengler JD, Koutrakis P, Ware JH, Raizenne M, Speizer
  2            FE. (1996). Health effects of acid aerosols on North American children: respiratory symptoms. Environ
  3            Health Perspect. 104(5):500-5.

  4    Dockery DW, Luttmann-Gibson H, Rich DQ, Link MS, Mittleman MA, Gold DR, Koutrakis P, Schwartz JD,
  5            Verrier RL (2005). Association of air pollution with increased incidence of ventricular tachyarrhythmias
  6            recorded by implanted cardioverter defibrillators. Environ Health Perspect, 113: 670-674.

  7    Dominici F, Peng RD, Bell ML, Pham L, McDermott A, Zeger SL, Samet JL. (2006a). Fine particulate air pollution
  8            and hospital admission for cardiovascular and respiratory diseases. JAMA, 295: 1127-1134.

  9    Dominic F. (2006b). Letter from Francesca Dominici, Associate Professor of Biostatistics, Bloomberg School of
10            Public Health, Johns Hopkins University, comments to the proposed rule.  DocketNo. OAR-2001-0017-
11            0988.  March 21, 2006.

12    Dominici F, Peng RD, Zeger SL, White RH, Samet JM. (2007).  Particulate air pollution and mortality in the United
13            States: did the risks change from 1987 to 2000? Am J Epidemiol, 166: 880-8.

14    Dominici F, Peng RD, Ebisu K, Zeger SL, Samet JM, Bell ML (2007). Does the effect of PM10 on mortality depend
15            on PM nickel and vanadium content? A reanaly sis of the NMMAPS data. Environ Health Perspect,  115:
16            1701-1703.

17    Dubowsky SD, Suh H, Schwartz J, Coull BA, Gold DR (2006). Diabetes, obesity, and hypertension may enhance
18            associations between air pollution and markers of systemic inflammation. Environ Health Perspect,  114:
19            992-998.

20    Eftim SE, Samet JM, Janes H, McDermott A, Dominici F. (2008). Fine Particulate Matter and Mortality: A
21            Comparison of the Six Cities and American Cancer Society Cohorts With a Medicare Cohort.
22            Epidemiology, 19: 209- 216.

23    Enstrom JE. (2005). Fine particulate air pollution and total mortality  among elderly Californians,  1973-2002. Inhal
24            Toxicol, 17: 803-816.

25    Fairley D (2003). Mortality and air pollution for Santa Clara County, California, 1989-1996, In: Revised analyses of
26            time-series studies of air pollution and health. Special report. Health Effects Institute. Boston, MA.
27            http://www.healtheffects.org/Pubs/TimeSeries.pdf.

28    Fakhri AA, Ilic LM, Wellenius GA, Urch B, Silverman F, Gold  DR,  Mittleman MA (2009). Autonomic effects of
29            controlled fine particulate exposure in young healthy adults: Effect modification by ozone. Environ Health
30            Perspect, 117: 1287-1292.

31    Gauderman WJ., McConnell R, Gilliland F, London S, Thomas D, Avol E, Vora H., Berhane K., Rappaport EB.,
32            Lurmann F, Margolis HG, Peters J.  (2000). Association between air pollution and lung function growth in
33            southern California children.  Am. J. Respir.  Crit. Care Med. 162: 1383-1390.

34    Gauderman WJ, Gilliland GF, Vora H, Avol E, Stram D, McConnell R, Thomas D, Lurmann F, Margolis HG,
3 5            Rappaport, E. B., Berhane, K., Peters, J. M. (2002). Association between air pollution and lung function
36            growth in southern California children: results from a second cohort. Am. J. Respir. Crit. Care Med. 166:
37            76-84.

38    Gauderman W.J, Avol, E., Gilliland, F., Vora, H., Thomas, D., Berhane, K., McConnell, R., Kuenzli, N., Lurmann,
39            F., Rappaport, E., Margolis, H., Bates, D., Peters, J.(2004). The effect of air pollution on lung development
40            from 10 to 18 years of age.  NEJM 351:1057-67.
              Draft Do Not Quote or Cite            2-111                           March 2010

-------
  1     Gent JF, Koutrakis P, Belanger K, Triche E, Holford TR, Bracken MB, Leaderer BP. (2009) Symptoms and
  2            medication use in children with asthma and traffic-related sources of fine particle pollution. Environ
  3            Health Perspect.  117:1168-74.

  4     Gold DR, Litonjua AA, Zanobetti A, Coull BA, Schwartz J, MacCallum G, Verrier RL, Nearing BD, Canner MJ,
  5            Suh H, Stone PH (2005). Air pollution and ST-segment depression in elderly subjects. Environ Health
  6            Perspect 113: 883-887.

  7     Gong H Jr, Linn WS, Clark KW, Anderson KR, Geller MD, Sioutas C. (2005) Respiratory responses to exposures
  8            with fine particulates and nitrogen dioxide in the elderly with and without COPD.  Inhal Toxicol.
  9            17(3): 123-32.

10     Goss CH: Newsom SA, Schildcrout JS, Sheppard L, Kaufman JD. (2004). Effect of ambient air pollution on
11            pulmonary exacerbations and lung function in cystic fibrosis. Am J Respir Crit Care Med, 169: 816-821.

12     Hassett-Sipple B and Stanek L (2009) Paniculate Matter Air Quality Data Requested from Epidemiologic Study
13            Authors. Memorandum to PM NAAQS review docket EPA-HQ-OAR-2007-0492.

14     HEI (2009). Health Effects Institute Health Review Committee Commentary for Research Report No. 140,
15            Extended Follow-up and Spatial Analysis of the American Cancer Society Study Linking Paniculate Air
16            Pollution and Mortality. HEI Research Report 140. Health Effects Institute, Boston, MA. pp. 115-136.
17            Available:  http://pubs.healtheffects.org/view.php?id=315.

18     Henderson, R. (2006a). Letter from Dr. Rogene Henderson, Chair, Clean Air Scientific Advisory Committee to the
19            Honorable Stephen L. Johnson, Administrator, US EPA.  Clean Air Scientific Advisory Committee
20            Recommendations Concerning the Proposed National Ambient Air Quality Standards for Paniculate
21            Matter. March 21, 2006. Available: http://www.epa.gov/sab/pdf/casac-ltr-06-002.pdf

22     Henderson R., Cowling, E., Crapo, ID.., Miller, F. J., Poirot, R.L., Speizer, F., Zielinska, B. (2006b).  Letter from
23            Clean Air Scientific Advisory Committee to the Honorable Stephen L. Johnson, Administrator, US EPA.
24            Clean Air Scientific Advisory Committee Recommendations Concerning the Final National Ambient Air
25            Quality Standards for Paniculate Matter. September 29, 2006. Available:
26            http://www.epa.gov/sab/pdf/casac-ltr-06-002.pdf

27     Hopke PK, Ito K, Mar T, Christensen WF, Eatough DJ, Henry RC, Kim E, Laden F, Lall R, Larson TV, Liu H, Neas
28            L, Pinto J, Stolzel M, Suh H, Paatero P, Thurston GD (2006). PM source apportionment and health effects:
29            1 Intercomparison of source apportionment results. J Expo Sci Environ Epidemiol, 16: 275-286.

30     Islam T, Gauderman WJ, Berhane K, McConnell R, Avol E, Peters JM, Gilliland FD. (2007). The relationship
31            between air pollution, lung function and asthma in adolescents. Thorax, 62: 957-963.

32     Ito K, Christensen WF, Eatough DJ, Henry RC, Kim E, Laden F, Lall R, Larson TV, Neas L, Hopke PK, Thurston
33            GD (2006). PM source apportionment and health effects: 2 An investigation of intermethod variability in
34            associations between source-apportioned fine particle mass and daily mortality in Washington, DC. J Expo
3 5            Sci Environ Epidemiol,  16: 300-310.

36     Jerret M, Burnett RT, Ma  R: Pope CA, Kreski D, Newbold KB, Thurston G, Shi Y, Finkelstein N, Calle N, Thun
37            MJ. (2005) Spatial analysis of air pollution and mortality in Los Angeles. Epidemiology, 16: 727-36.

3 8     Kim JJ, Smorodinsky S, Lipsett M, Singer BC, Hodgson AT, Ostro B. (2004). Traffic-related air pollution near busy
39            roads: the East Bay children's Respiratory Health Study. Am J Respir Crit Care Med, 170: 520-526.

40     Krewski, D., Burnett, R. T., Goldberg, M. S., Hoover, K., Siemiatycki, J,,  Jerrett, M., Abrahamowicz, M., White, W.
41            H. (2000). Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of
              Draft Do Not Quote or Cite            2-112                           March 2010

-------
  1            paniculate air pollution and mortality. A special report of the Institute's particle epidemiology reanalysis
  2            project. Cambridge, MA: Health Effects Institute. http://pubs.healtheffects.org/view.php?id=6

  3     Krewski D, Jerrett M, Burnett RT, Ma R, Hughes E, Shi Y, Turner MC, Pope AC III, Thurston G, Calle EE, Thun
  4            MJ. (2009). Extended Follow-Up and Spatial Analysis of the American Cancer Society Study Linking
  5            Paniculate Air Pollution and Mortality. HEI Research Report 140, Health Effects Institute, Boston, MA.
  6            Available: http://pubs.heartheffects.org/view.php?id=315.

  7     Kunzli N, Medina S, Kaiser R, et al. (2001). Assessment of deaths attributable to air pollution: should we use risk
  8            estimates based on time series or on cohort studies? Am J Epidemiol 153:1050 -1055.

  9     Laden F, Neas LM, Dockery DW, Schwartz J. (2000). Association of fine paniculate matter from different sources
10            with daily mortality in six US cities. Environ Health Perspect, 108: 941-947.

11     Laden F, Schwartz J,  Speizer FE, Dockery DW. (2006). Reduction in fine paniculate air pollution and mortality:
12            extended follow-up of the Harvard Six Cities Study. Am. J. Respir. Crit. Care. Med. 173: 667-672.

13     Laden F. (2009). Personal communication with Dr. Francine Laden: Annual PM2 5 levels used in the update of the
14            Harvard Six Cities Study. May 21, 2009. Docket EPA-HQ-OAR-2007-0492

15     Lanki T, Hoek G, Timonen K, Peters A, Tiittanen P, Vanninen E, Pekkanen J (2008). Hourly variation in fine
16            particle exposure is associated with transiently increased risk of ST segment depression. Br Med J, 65: 782-
17            786.

18     Lin M, Stieb DM, Chen Y (2005). Coarse paniculate matter and hospitalization for respiratory infections in children
19            younger than 15 years  in Toronto: a case-crossover analysis. Pediatrics, 116: 235-240.

20     Lipfert FW, Baty JD, Miller JP, Wyzga RE. (2006). PM2 5 constituents and related air quality variables as predictors
21            of survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657.

22     Lippmann M, Ito K, Hwang JS, Maciejczyk P, Chen LC (2006). Cardiovascular effects of nickel in ambient air.
23            Environ Health Perspect, 114: 1662-1669.

24     Lisabeth LD, Escobar JD, Dvonch JT, Sanchez BN, Majersik JJ, Brown DL,  Smith MA, Morgenstern LB (2008).
25            Ambient air pollution and risk for ischemic stroke and transient ischemic attack. Ann Neurol, 64: 53-59.

26     Ljungman PLS, Berglind N, Holmgren C, Gadler F, Edvardsson N, Pershagen G, Rosenqvist M, Sjogren B,
27            Bellander T  (2008). Rapid effects of air pollution on ventricular arrhythmias. Eur Heart J, 29: 2894-2901.

28     Lund AK, Lucero J, Lucas S, Madden MC, McDonald JD, Seagrave JC, Knuckles TL, Campen MJ (2009).
29            Vehicular emissions induce vascular MMP-9 expression and activity associated with endothelin-1 mediated
30            pathways. Arterioscler Thromb Vase Biol, 29: 511-517.

31     Mar TF., Norris GA., Larson TV., Wilson, W. E., Koenig, J.  Q. (2003) Air pollution and cardiovascular mortality in
32            Phoenix, 1995-1997. In: Revised analyses of time-series studies of air pollution and health. Special report.
33            Boston, MA: Health Effects Institute, pp. 177-182. Available: http://www.healtheffects.org/news.htm [16
34            May, 2003].

3 5     Mar TF, Larson TV, Stier RA, Claiborn C, Koenig JQ (2004). An analysis of the association between respiratory
36            symptoms in subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol, 16: 809-
37            815.

3 8     Mar TF, Ito K, Koenig JQ, Larson TV, Eatough DJ, Henry RC, Kim E, Laden F, Lall R, Neas L,  Stolzel M, Paatero
39            P, Hopke PK, Thurston GD.(2006) PM source apportionment and health effects. 3. Investigation of inter-
              Draft Do Not Quote or Cite           2-113                            March 2010

-------
  1            method variations in associations between estimated source contributions of PM2.5 and daily mortality in
  2            Phoenix, AZ. J Expo Sci Environ Epidemiol. 16:311-20

  3    Mar TF, Jansen K, Shepherd K, Lumley T, Larson TV, Koenig JQ (2005). Exhaled nitric oxide in children with
  4            asthma and short-term PM25 exposure in Seattle. Environ Health Perspect, 113: 1791-1794.

  5    McConnell R, Berhane K, Gilliland F, Molitor J, Thomas D, Lurmann F, Avol E, Gauderman WJ, Peters JM.
  6            (2003). Prospective study of air pollution and bronchitic symptoms in children with asthma.  Am J Respir
  7            Crit Care Med, 168: 790-797.

  8    Metzger KB, Tolbert PE, Klein M,  Peel JL, Flanders WD, Todd KH, Mulholland JA, Ryan PB, Frumkin H. (2004).
  9            Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15: 46-56.

10    Metzger KB, Klein M, Flanders WD, Peel JL, Mulholland JA, Langberg JJ, Tolbert PE. (2007). Ambient air
11            pollution and cardiac arrhythmias in patients with implantable defibrillators. Epidemiology, 18:  585-592.

12    Miller KA, Siscovick DS, Sheppard L, Shepherd K, Sullivan JH, Anderson GL, Kaufman JD. (2007). Long-term
13            exposure to air pollution and incidence of cardiovascular events in women. N Engl J Med, 356: 447-458.

14    Ostro B.D., Broadwin, R., Green, S., Feng, W., Lipsett, M.(2006). Fine paniculate air pollution and mortality in
15            nine California counties: results from CALFINE.  Env. Health Perspect. 114: 29-33.

16    Peel JL, Tolbert PE, Klein M, Metzger KB, Flanders WD, Knox T, Mulholland JA, Ryan PB, Frumkin H. (2005).
17            Ambient air pollution and respiratory emergency department visits. Epidemiology, 16: 164-174.

18    Peel JL, Metzger KB, Klein M, Flanders WD, Mulholland JA, Tolbert PE. (2007). Ambient air pollution and
19            cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165: 625-633.

20    Peng RD, Bell ML, Geyh AS, McDermott A, Zeger SL, Samet JM, Dominici F (2009). Emergency admissions for
21            cardiovascular and respiratory diseases and the chemical composition of fine particle air pollution. Environ
22            Health Perspect. 117:957-63.

23    Penttinen P, Vallius M, Tiittanen P, Ruuskanen J, Pekkanen J (2006). Source-specific fine particles in urban air and
24            respiratory function among adult asthmatics. Inhal Toxicol, 18: 191-198.

25    Peretz A, Kaufman JD, Trenga CA, Allen J, Carlsten C, Aulet MR, Adar SD, Sullivan JH (2008). Effects of diesel
26            exhaust inhalation on heart rate variability in human volunteers. Environ Res, 107: 178-184.

27    Peters A., Dockery, D. W., Muller, J.  E., Mittleman, M. A. (2001). Increased paniculate air pollution and the
28            triggering of myocardial infarction. Circulation 103:2810-2815.

29    Pope CA 3rd, Thun, M.J., Namboodiri, M.M., Dockery, D.W., Evans, J.S., Speizer, F.E., Heath,, C.W. (1995).
3 0            Paniculate air pollution as a  predictor of mortality in a prospective study of U. S. adults. Am J Respir Crit
31            Care Med.  151:669-674.

32    Pope CA 3rd, Burnett, R. T., Thun, M. J., Calle, E. E., Krewski, D., Ito, K., Thurston, G. D. (2002).  Lung cancer,
33            cardiopulmonary mortality, and long-term exposure to fine paniculate air pollution. J. Am. Med. Assoc.
34            287:1132-1141.

3 5    Pope CA 3rd, Burnett RT, Thurston GD, Thun MJ, Calle EE, Krewski D, Godleski JJ (2004). Cardiovascular
36            mortality and long-term exposure to paniculate air pollution: epidemiological evidence of general
37            pathophysiological pathways of disease. Circulation, 109: 71-77.

3 8    Pope CA 3rd, Ezzati M, Dockery DW. (2009). Fine-particulate air pollution and life expectancy in the United States.
39            N Engl J Med, 360: 376-386.

              Draft Do Not Quote or Cite            2-114                           March 2010

-------
  1    Rabinovitch N, Zhang LN, Murphy JR, Vedal S, Dutton SJ, Gelfand EW (2004). Effects of wintertime ambient air
  2            pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J Allergy
  3            Clinlmmunol, 114: 1131-1137.

  4    Rabinovitch N, Strand M, Gelfand EW (2006). Paniculate levels are associated with early asthma worsening in
  5            children with persistent disease. Am J Respir Crit Care Med, 173: 1098-1105.

  6    Rabl A. (2003) Interpretation of air pollution mortality: number of deaths or years of life lost? J Air Waste Manag
  7            Assoc., 53:41-50.

  8    Raizenne M, Neas LM, Damokosh AI, Dockery DW, Spengler JD, Koutrakis P, Ware JH, Speizer FE. (1996).
  9            Health effects of acid aerosols on North American children: pulmonary function. Environ Health Perspect,
10            104:506-514.

11    Rich DQ, Schwartz J, Mittleman MA, Link M, Luttmann-Gibson H, Catalano PJ, Speizer FE, Dockery DW (2005).
12            Association of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J
13            Epidemiol, 161: 1123-1132.

14    Riediker M, Cascio WE, Griggs TR, Herbst MC, Bromberg PA, Neas L, Williams RW, Devlin RB (2004).
15            Paniculate matter exposure in cars is associated with cardiovascular effects in healthy young men. Am J
16            Respir Crit Care Med,  169: 934-940.

17    Rosenthal FS,  Carney JP, Olinger ML  (2008). Out-of-hospital cardiac arrest and airborne fine paniculate matter: a
18            case-crossover analysis of emergency medical services data in Indianapolis, Indiana. Environ Health
19            Perspect, 116:631-636.

20    Ross M, Langstaff, J. (2005). Updated statistical information on air quality data from epidemiologic studies.
21            Memorandum to PMNAAQS review docket EPA-HQ-OAR-2001-0017. Docket No. OAR-2001-0017-
22            0261. January 31,2005.

23    Ross M, Langstaff, J. (2006). Statistical information on air quality data from additional epidemiologic studies.
24            Memorandum to PM NAAQS review docket EPA-HQ-OAR-2001-0017. Docket ID No. OAR-2001-0017-
25            1409. April 5, 2006.

26    Rundell KW, Hoffman JR, Caviston R, Bulbulian R, Hollenbach AM (2007). Inhalation of ultrafine and fine
27            paniculate matter disrupts systemic vascular function. Inhal Toxicol, 19: 133-140.

28    Rundell KW, Caviston R (2008). Ultrafine and fine paniculate matter inhalation decreases exercise performance in
29            healthy subjects. J Strength Cond Res, 22: 2-5.

30    Schmidt M, Frank, N., Mintz, D., Rao, T., McCluney, L. (2005). Analyses of paniculate matter (PM) data for the
31            PM NAAQS review. Memorandum to PMNAAQS review docket EPA-HQ-OAR-2001-0017. June 30,
32            2005.

33    Schmidt M (2010). Analysesof PM2.5 data for the PMNAAQS review. Memorandum to PM NAAQS review
34            docket EPA-HQ-OAR-2007-0492.

3 5    Schwartz J (2003). Daily deaths associated with air pollution in six US cities and short-term mortality displacement
36            in Boston. From Revised Analyses of Time-Series Studies  of Air Pollution and Health, Health Effects
3 7            Institute Special Report, May 2003.

3 8    Schwartz J, Coull B, Laden F, Ryan L  (2008). The effect of dose and timing of dose on the association between
39            airborne particles and survival. Environ Health Perspect, 116: 64-69.
              Draft Do Not Quote or Cite            2-115                           March 2010

-------
  1     Shah AP, Pietropaoli AP, Frasier LM, Speers DM, Chalupa DC, Delehanty JM, Huang LS, Utell MJ, Frampton MW
  2            (2008). Effect of inhaled carbon ultrafine particles on reactive hyperemia in healthy human subjects.
  3            Environ Health Perspect,  116: 375-380.

  4     Slaughter JC, Lumley T, Sheppard L, Koenig JQ, Shapiro GG. (2003). Effects of ambient air pollution on symptom
  5            severity and medication use in children with asthma. Ann Allergy Asthma Immunol, 91: 346-353.

  6     Slaughter JC, Kim E, Sheppard L,  Sullivan JH, Larson TV, Claiborn C (2005). Association between paniculate
  7            matter and emergency room visits, hospital admissions and mortality in Spokane, Washington. J Expo Sci
  8            Environ Epidemiol, 15: 153-159.

  9     Stieb DM, Beveridge RC, Brook JR, Smith-Doiron M, Burnett RT, Dales RE, Beaulieu S, Judek S, Mamedov A
10            (2000). Air pollution, aeroallergens and cardiorespiratory emergency department visits in Saint John,
11            Canada. J Expo Sci Environ Epidemiol, 10: 461-477.

12     Sullivan J, Sheppard L, Schreuder A, Ishikawa N, Siscovick D, Kaufman J (2005). Relation between short-term
13            fine-particulate matter exposure and onset of myocardial infarction. Epidemiology,  16: 41-48.

14     Thurston G, Ito K, Mar T, Christensen WF, Eatough D J, Henry RC, Kim E, Laden F, Lall R, Larson TV, Liu H,
15            Neas L, Pinto J, Stolzel M, Suh H, Hopke PK (2005). Results and implications of the workshop on the
16            source apportionment of PM health effects. Epidemiology,  16: S134-S135.

17     Tolbert PE, Klein M, Peel JL, Sarnat SE, Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air
18            quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 24: 938-945.

19     Tsai FC, Apte MG, Daisey JM (2000). An exploratory analysis of the relationship between mortality and the
20            chemical composition of airborne paniculate matter. Inhal Toxicol, 12: 121-135.

21     Urch B, Silverman F, Corey P, Brook JR, Lukic KZ, Rajagopalan S, Brook RD (2005). Acute blood pressure
22            responses in healthy adults during controlled air pollution exposures. Environ Health Perspect, 113: 1052-
23            1055.

24     US EPA (1996a). Air Quality Criteria for Paniculate Matter. Research Triangle Park, NC: National Center for
25            Environmental Assessment-RTF Office, report no. EPA/600/P-95/001. 3v.  Available:
26            http://www.epa.gov/ttn/naaqs/standards/pm/sjmjr cd.html.

27     US EPA (1996b). Review of the National Ambient Air Quality Standards for Paniculate Matter: Policy Assessment
28            of Scientific and Technical Information, OAQPS Staff Paper. Research Triangle Park, NC 27711: Office
29            of Air Quality Planning and Standards, report no. EPA-452YR-96-013. Available:
30            http://www.epa.gov/ttn/naaqs/standards/pm/sjmjr sp.html.

31     US EPA (2004).  Air Quality Criteria for Paniculate Matter. National Center for Environmental Assessment, Office
32            of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711,
33            report no. EPA/600/P-99/002aF and EPA/600/P-99/002bF. October 2004. Available:
34            http://www.epa.gov/ttn/naaqs/standards/pm/sjm crcd.html

35     US EPA (2005). Review of the National Ambient Air Quality Standards for Paniculate Matter: Policy Assessment
36            of Scientific and Technical Information, OAQPS Staff Paper. Research Triangle Park, NC 27711: Office
37            of Air Quality Planning and Standards, report no. EPA EPA-452/R-05-005a. December 2005.
38                    Available: http://www.epa.gov/ttn/naaqs/standards/pm/sjm cr sp.html

39     US EPA (2008a). Integrated Review Plan for the National Ambient Air Quality Standards fro Paniculate Matter.
40            National Center for Environmental Assessment and Office of Air Quality Planning and Standards, U.S.
41            Environmental Protection Agency, Research Triangle Park, NC. Report No. EPA 452/R-08-004. March
              Draft Do Not Quote or Cite           2-116                           March 2010

-------
  1            2008.  Available at:
  2            http://www.epa.gov/ttn/naaqs/standards/pm/data/2008 03 final  integrated review_plan.pdf

  3     US EPA (2009a). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
  4            Protection Agency, Washington, DC, EPA/600/R-08/139F, December 2009. Available:
  5            http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=216546

  6     US EPA (2010a). Quantitative Risk Assessment for Paniculate Matter - Second External Review Draft. Office of
  7            Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC.
  8            EPA-452/P-09-001. February 2010.  Available:
  9            http://www.epa.gov/ttn/naaqs/standards/pm/sjm index.html.

10     Villeneuve PJ, Chen L, Stieb D, Rowe BH (2006). Associations between outdoor air pollution and emergency
11            department visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21: 689-700.

12     Woodruff TJ, Darrow LA, Parker JD (2008). Air pollution and postneonatal infant mortality in the United States,
13            1999-2002. Environ Health Perspect, 116:  110-115.

14     Zanobetti A, Schwartz J. (2009). The effect of fine and coarse paniculate air pollution on mortality: A national
15            analysis. Environ Health Perspect, 117: 1-40.

16     Zhang Z, Whitsel E, Quibrera P, Smith R, Liao D, Anderson G, Prineas R (2009). Ambient fine paniculate matter
17            exposure and myocardial ischemia in the Environmental Epidemiology of Arrhythmogenesis in the
18            Women's Health Initiative (EEAWHI) study. Environ Health Perspect, 117: 751-756.

19     Zeger S, McDermott A, Dominicit F, Samet J. (2007). Mortality in the Medicare population and chronic exposure
20            to fine paniculate air pollution. Johns Hopkins University. Baltimore.
21            http://www.bepress.com/ihubiostat/paperl33.

22     Zeger S, Dominici F, McDermott A, Samet J. (2008). Mortality in the Medicare population and chronic exposure to
23            fine paniculate air pollution in urban centers (2000-2005). Environ Health Perspect, 116: 1614.
              Draft Do Not Quote or Cite           2-117                            March 2010

-------
 1             3   REVIEW OF THE PRIMARY STANDARD FOR THORACIC
 2                                       COARSE PARTICLES

 3           This chapter presents preliminary staff conclusions with regard to the adequacy of the
 4    current primary PMio standard, which is intended to protect public health against exposures to
 5    thoracic coarse particles (PMio-2.s), and potential alternative primary standards for consideration
 6    in this review. Our assessment of these issues is framed by a series of key policy-relevant
 7    questions, which expand upon those presented in the IRP (US EPA, 2008a) at the outset of this
 8    review.  The answers to these questions will inform decisions on whether, and if so how, to
 9    revise the current PMio standard.
10           Staff notes that final decisions regarding the primary PMio standard will draw upon
11    scientific information about health effects, as well as judgments about how to deal with the range
12    of uncertainties that are inherent in the scientific evidence. Ultimately, the final decisions are
13    largely public health policy judgments. Our approach to informing these judgments, discussed
14    more fully below, recognizes that the available health effects evidence reflects a continuum
15    consisting of ambient levels at which scientists generally agree that health effects are likely to
16    occur through lower levels at which the likelihood and magnitude of the response become
17    increasingly uncertain
18           Our current approach for reviewing the primary PMio standard is presented in section 3.1.
19    Our preliminary conclusions regarding the adequacy of the current PMio standard are presented
20    in section 3.2.  Section 3.3 presents our preliminary considerations and conclusions with respect
21    to potential alternative standards, focusing on each of the basic elements of the standards:
22    pollutant indicator (section 3.3.1), averaging time (section  3.3.2), and form and level (section
23    3.3.3).  The next draft of this chapter will also include an initial overview of key uncertainties
24    and suggested future research areas and data collection efforts.

25    3.1     APPROACH
26           Staffs approach for reviewing the current primary  PMio standard builds upon the
27    approaches used  in previous PM NAAQS reviews.  Our current approach is based on the updated
28    scientific and technical information in the ISA.1  The past and current approaches described
29    below are all based most fundamentally on using information from epidemiologic studies to
30    inform the selection of PM standards that, in the Administrator's judgment, protect public health
31    with an adequate margin of safety. In the case of thoracic  coarse  particles, such information is in
      1 As described below, given limitations in the health effects data and in the monitoring network, we have not
      conducted a quantitative risk assessment in this review for thoracic coarse particles.
            Draft Do Not Quote or Cite            3-1                         March 2010

-------
 1    the form of PMio-2.5 and PMio air quality distributions over which health effect associations have
 2    been reported. In light of limitations in PMio-2.5 epidemiologic studies, dosimetric and
 3    toxicological information also play an important role, especially in considerations related to the
 4    appropriate indicator for a standard meant to protect against exposures to thoracic coarse
 5    particles. Evidence-based approaches to using information from epidemiologic studies to inform
 6    decisions on PM standards are complicated by the recognition that no population threshold,
 7    below which it can be concluded with confidence that PM-related  effects do not occur, can be
 8    discerned from the available evidence.  As a result, any approach to reaching decisions on what
 9    standards are appropriate requires judgments about how to translate the information available
10    from the epidemiologic studies into a basis for appropriate standards, which includes
11    consideration of how to weigh the uncertainties in reported associations across the distributions
12    of PM concentrations in the studies.  Such approaches are consistent with setting standards that
13    are neither more nor less stringent than necessary, recognizing that a zero-risk standard is not
14    required by the CAA.

15    3.1.1  Approaches Used in Previous Reviews
16              3.1.1.1  Reviews Completed in 1987 and 1997
17          The PM NAAQS have always included some type of a primary standard to protect
18    against effects associated with exposures to thoracic coarse particles.  In 1987, when EPA first
19    revised the PM NAAQS, EPA changed the indicator for PM from  Total Suspended Particles
20    (essentially applicable to particles  smaller than 25-45 micrometers) to focus on inhalable
21    particles, those which can penetrate into the trachea, bronchi, and deep lungs (52 FR 24634).
22    EPA changed the PM indicator to PMio based on evidence that the risk of adverse health effects
23    associated with particles of 10 micrometers or less was  significantly greater than that associated
24    with larger particles (52 FR at 24639).
25          In the  1997 review, in conjunction with establishing new fine particle (i.e., PM2.s)
26    standards (see above, sections 1.2.3,2.1.1), EPA concluded that continued protection remained
27    warranted against potential effects associated with thoracic coarse particles in the size range of
28    2.5 to 10 |im.  This conclusion was based on particle dosimetry, toxicological information, and
29    on limited epidemiologic evidence from studies that measured PMio in areas where coarse
30    particles were likely to dominate the distribution (62 FR 38677, July 18, 1997). Thus, EPA
31    concluded that the PMio standards would provide protection against effects associated with
32    particles in the narrower size range of 2.5 to 10 |im.  Although  a more narrowly defined indicator
33    was considered in that review (i.e., PMio-2.s), EPA concluded that it was more appropriate, based
34    on existing evidence, to continue to use PMio as the indicator for standards to control thoracic
            Draft Do Not Quote or Cite           3-2                         March 2010

-------
 1    coarse particles. This decision was based, in part, on the recognition that the only studies of
 2    clear quantitative relevance to health effects most likely associated with thoracic coarse particles
 3    used PMio in areas where the coarse fraction was the dominant fraction of PMi0, namely two
 4    studies conducted in areas that substantially exceeded the 24-hour PMio standard (62 FR 38679).
 5    In addition, this decision reflected the fact that there were only very  limited ambient air quality
 6    data then available specifically for thoracic coarse particles, in contrast to the extensive
 7    monitoring network already in place for PMio, Therefore, it was more administratively feasible
 8    to use PMio as an indicator. EPA also stated that the PMio  standards would work in conjunction
 9    with the PM2.5 standards by regulating the portion of particulate pollution not regulated by the
10    PM2.5 standards.2

11                3.1.1.2   Review Completed in 2006
12            In the review of the PM NAAQS that concluded in 2006, EPA considered the growing,
13    but still limited, body of evidence supporting associations between health effects and thoracic
14    coarse particles measured as PMio-2.5.3 The new studies available in the 2006 review included
15    epidemiologic studies that reported associations with health effects using direct measurements of
16    PMio-2.5, as well as dosimetric and toxicological studies. In light of this growing body of
17    evidence, staff concluded that it was appropriate to revise the PMio standards and to base any
18    revised standards principally on available evidence and air quality information for PMio-2.5-  Staff
19    also concluded that it was appropriate to consider evidence from studies that measured PMio in
20    locations where the majority of PMio was in the PMio-2.5 fraction (US EPA, 2005, section 5.4.1)
21    and that the level of protection afforded by the existing 1987 PMio standard remained
       As explained in chapter 1, in May 1998, a three-judge panel of the U.S. Court of Appeals for the District of
      Columbia Circuit found "ample support" for EPA's decision to regulate coarse particle pollution, but vacated the
      1997 PM10 standards, concluding that EPA had failed to adequately explain its choice of PM10 as the indicator for
      thoracic coarse particles pointing to the lack of reasoned explanation for the variable level of allowable
      concentrations of thoracic coarse particles (varying by levels of PM2 5) and the consequent double regulation of
      PM2.5. American Trucking Associations v. EPA , 175 F. 3d 1027, 1054-56 (D.C. Cir. 1999).. The court also rejected
      considerations of administrative feasibility as justification for a NAAQS, which are based exclusively on health and
      welfare considerations. IxL at 1054. Pursuant to the  court's decision, EPA removed the vacated 1997 PM10
      standards from the Code of Federal Regulations (CFR) (69 FR 45592, July 30, 2004) and deleted the regulatory
      provision [at 40 CFR section 50.6(d)] that controlled the transition from the pre-existing 1987 PM10 standards to the
      1997 PM10 standards (65 FR 80776, December 22, 2000). The pre-existing 1987 PM10 standards remained in place.
      Id. at 80777.  As noted in chapter 1 and in more detail above, in the 2006 review of the PM NAAQS, EPA re-
      addressed the appropriateness of using PM10 as an indicator for a standard meant to protect against exposures to
      thoracic coarse particles, and the decision to use PM10 as the indicator was upheld on judicial review. American
      Farm Bureau Federation v. EPA. 559 F. 3d 512, 533-38 (D.C. Cir. 2009).
      3ThePM Staff Paper (US EPA, 2005) also presents results of a quantitative assessment of health risks for PM10.2.5.
      However, staff concluded that the nature and magnitude of the uncertainties and concerns associated with this risk
      assessment weighed against its use as a basis for recommending specific levels for a thoracic coarse particle
      standard.
             Draft Do Not Quote or Cite             3-3                           March 2010

-------
 1    appropriate (US EPA, 2005, p. 5-67).  Responding to CASAC advice, staff also recommended
 2    that the indicator for thoracic coarse particles be urban coarse particles in the size range of 10-2.5
 3    micrometer range, thus focusing on those thoracic coarse particles that are generally present in
 4    urban environments (US EPA, 2005, p. 5-71). The agency proposed to retain a standard for a
 5    subset of thoracic coarse particles, proposing a qualified PMio-2.5 indicator to focus on the mix of
 6    thoracic coarse particles generally present in urban environments.  More specifically, the revised
 7    thoracic coarse particle standard would have applied only to an ambient mix of PMi0-2.5 that is
 8    dominated by resuspended dust from high-density traffic on paved roads and/or by industrial and
 9    construction sources. The revised standard would not have applied to any ambient mix of PMio-
10    2.5 that is dominated by rural windblown dust and soils. In addition, agricultural sources, mining
11    sources, and other similar sources of crustal material would not have been subject to control in
12    meeting the standard (71 FR 2667 to 2668, January 17, 2006).
13          The Agency received a large number of comments that were overwhelmingly opposed to
14    the proposed qualified PMio-2.5 indicator (71 FR 61188 to 61197).  After careful consideration of
15    the scientific evidence and the recommendations contained in the 2005 Staff Paper, the advice
16    and recommendations from CASAC, the public comments received regarding the appropriate
17    indicator for coarse particles, and after extensive evaluation of the alternatives available to the
18    Agency, the Administrator decided it would not be appropriate to adopt a qualified PMio-2.5
19    indicator. Underlying this determination was the decision that it was  requisite to provide
20    protection from exposure to all thoracic coarse PM, regardless of its origin, rejecting arguments
21    that there are no health effects from community-level exposures to coarse PM in non-urban areas
22    (71 FR 61189).  The EPA concluded that dosimetric, lexicological, occupational and
23    epidemiologic evidence supported retention of a primary standard for short-term exposures that
24    included all thoracic coarse particles (i.e. both urban and non-urban),  consistent with the Act's
25    requirement that primary NAAQS provide an adequate margin of safety. Given the serious
26    potential effects, the large numbers of persons exposed, and the need  for a standard to provide an
27    adequate margin of safety, the agency  concluded it was appropriate to retain a standard applying
28    to all coarse particles (71 FR 61197). At the same time, the agency  concluded that the standard
29    should target protection toward urban  areas, where the evidence of health effects from exposure
30    to PMio-2.5 was strongest (71 FR at 61193, 61197). The proposed indicator was not suitable for
31    that purpose.  Not only did it inappropriately provide no protection at all to many areas, but it
32    failed to identify many areas where the ambient mix was dominated by coarse particles
33    contaminated with urban/industrial types of coarse particles for which evidence of health effects
34    was strongest (71 FR 61193).

            Draft Do Not Quote or Cite           3-4                         March 2010

-------
 1          The agency ultimately concluded that the existing indicator, PMio, was most consistent
 2    with the evidence.  Although PMio includes both coarse and fine PM, it remained an appropriate
 3    indicator for thoracic coarse particles because fine particle levels are generally higher in urban
 4    areas and, therefore, a PMio standard set at a single unvarying level will generally result in lower
 5    allowable concentrations of thoracic coarse particles in urban areas than in non-urban areas.
 6    This was considered to be an appropriate targeting of protection given that the strongest evidence
 7    for effects associated with thoracic coarse particles came from epidemiologic studies conducted
 8    in urban areas and that elevated fine particle  concentrations in urban areas could result in
 9    increased contamination of coarse fraction particles by PM2.5, potentially  increasing the toxicity
10    of thoracic coarse particles in urban areas (71 FR 61195-96). Given the evidence that the
11    existing PMio standard afforded requisite protection with an ample margin of safety, the Agency
12    retained the level and form of the 24-hour standard.4
13          The Agency also revoked the annual PMio standard, in light of the conclusion in the PM
14    Criteria Document  (US EPA, 2004) that the available evidence  does not suggest an association
15    with long-term exposure to PMi0-2.5 and the conclusion in the Staff Paper  (US EPA, 2005) that
16    there is no quantitative evidence that directly supports an annual standard.
17          In the same rulemaking, EPA also included a new Federal Reference Method (FRM) for
18    the measurement of PMio-2.s in the ambient air (71 FR 61212 to 61213). Although the standards
19    for thoracic coarse  particles do not use a PMi0-2.5 indicator, the new FRM for PMi0-2.5 was
20    established to provide a basis for approving Federal Equivalent Methods (FEMs) and to promote
21    the gathering of scientific data to support future reviews of the PM NAAQS.

22    3.1.2  Litigation of 2006 Final Rule for thoracic coarse particles
23          A number of groups filed suit in response to the final decisions made in the 2006 review.
24    See American Farm Bureau Federation and the National Pork Producers Council v. EPA (DC
25    Cir. 2009).  Among the petitions for review were challenges from industry groups on the
26    decision to retain the PMio indicator and the  level of the PMio standard and from environmental
27    and public health groups on the decision to revoke the annual PMio standard. The court upheld
28    both the daily PMio standard and the  decision to revoke the annual standard.
29          First, the court upheld EPA's  decision for a standard to cover all thoracic coarse PM,,
30    both of urban and non-urban origin. The court rejected arguments that the evidence showed
31    there are no risks from exposure to non-urban coarse PM. The court further found that EPA had
32    a reasonable basis not to set separate  standards for urban and non-urban coarse PM, namely the
      4 Thus, the level of 150 ug/m3 is met when this level is not exceeded more than once per year on average over a
      three-year period measured at each monitor within an area.
            Draft Do Not Quote or Cite           3-5                         March 2010

-------
 1    inability to reasonably define what ambient mixes would be included under either 'urban' or
 2    'non-urban;' and that the evidence in the record supported EPA's cautious decision to provide
 3    "some protection from exposure to thoracic coarse particles... in all areas."  559 F. 3d at 532-33.
 4    Specifically, the court stated,
 5
 6           Although the evidence of danger from coarse PM is, as EPA recognizes,
 7           "inconclusive," (71 FR 61193, October 17, 2006), the agency need not wait for
 8           conclusive findings before regulating a pollutant it reasonably believes may pose
 9           a significant risk to public health. The evidence in the record supports the EPA's
10           cautious decision that "some protection from exposure to thoracic coarse particles
11           is warranted in all areas." Id. As the court has consistently reaffirmed, the CAA
12           permits the Administrator to "err on the side of caution" in setting NAAQS.
13           559 F. 3d at 533.
14
15           The court also upheld EPA's decision to retain the level  of the standard at 150 |ig/m3 and
16    to use PMio as the indicator for a standard meant to protect against exposures to thoracic coarse
17    particles. In upholding the level of the standard, the court referred to the conclusion in the Staff
18    Paper that there is "little basis for concluding that the degree of protection afforded by the
19    current PMio standards in urban areas is greater than warranted, since potential mortality effects
20    have been associated with air quality levels not allowed by the current 24-hour standard, but
21    have not been associated with air quality levels that would generally meet that standard, and
22    morbidity  effects have been associated with  air quality levels that exceeded the current 24-hour
23    standard only a few times." 559 F.  3d at 534. The court also rejected arguments that a PMio
24    standard established at an unvarying level will result in arbitrarily varying levels of protection
25    given that the level of coarse PM would vary based on the amount of fine PM present. The court
26    agreed that the variation in allowable coarse PM accorded with the strength of the evidence:
27    typically less coarse PM would be allowed in urban areas (where levels of fine PM are typically
28    higher), in accord with the strongest evidence of health effects from coarse particles. 559 F. 3d
29    at 535-36. In addition, such regulation would not impermissibly double regulate fine particles,
30    since any additional regulation of fine particles (beyond that afforded by the primary PM2 5
31    standard) would be for a different purpose; to prevent contamination of coarse particles by fine
32    particles.  559 F. 3d at 535, 536. These same explanations explained the choice of PMio as an
33    indicator, and provided the reasoned explanation for that choice lacking in the record for the
34    1997 standard. 559 F. 3d at 536.
35           With regard to the challenge from environmental and public health groups, the court
36    upheld  EPA's decision to revoke the annual  PMio standard. Specifically, the court stated,

             Draft Do Not Quote or Cite           3-6                        March 2010

-------
 1           The EPA reasonably decided that an annual coarse PM standard is not necessary
 2           because, as the Criteria Document and the Staff Paper make clear, the latest
 3           scientific data do not indicate that long-term exposure to coarse particles poses a
 4           health risk. The CAS AC also agreed that an annual coarse PM standard is
 5           unnecessary.
 6           559 F. 3d at 538-39.
 7
 8    3.1.3   Approach in the Current Review
 9           The staffs approach in this review is consistent with the approaches ultimately taken in
10    previous reviews.  We have taken into account evidence-based considerations, including
11    consideration of the uncertainties associated with the evidence, to inform our preliminary
12    conclusions related to the adequacy of the current PMio standard and potential alternative
13    standards. In so doing, we are seeking to provide as broad an array of options as is supportable
14    by the available evidence, recognizing that the selection of a  specific approach to reaching final
15    decisions on the primary PMio standard will reflect the judgments of the Administrator as to
16    what weight to place on different aspects of the evidence and associated uncertainties. As
17    discussed in more detail in the Quantitative Health Risk Assessment for Paniculate Matter -
18    Second External Review Draft (second draft RA, US EPA, 2010a), we have not conducted a
19    quantitative assessment of health risks associated with PMio-2.5-  Staff concluded that limitations
20    in the monitoring network and in the health studies that rely on that monitoring network, which
21    would be the basis for characterizing PMio-2.5 exposures and risks, would introduce significant
22    uncertainty into a PMio-2.5 risk assessment such that the risk estimates generated would be of
23    limited utility in informing review of the standard. Therefore, staff concluded in the second draft
24    RA that a quantitative risk assessment for PMio-2.5 is not supportable at this time (US EPA,
25    2010a,  p. 2-6).
26           For the purposes of this first draft Policy Assessment  (PA), we have drawn from the
27    assessment and integration of the studies evaluated in the Integrated Science Assessment for
28    Particulate Matter (FinalReport) (ISA,  US EPA, 2009a).  The  discussions presented in this
29    chapter consider evidence from epidemiologic studies, controlled human  exposure studies, and
30    toxicological studies evaluating short- or long-term exposures to thoracic coarse particles; as
31    discussed in chapters 6 and 7, respectively, with supporting information related to dosimetry and
32    potential mode of action  (MOA) evidence as presented in chapters 4 and  5, respectively, as well
33    as the integration of evidence across disciplines presented in  chapter 2 of the ISA.
34           With respect to these  evidence-based considerations, we have considered causal
35    inferences identified in the ISA based on consideration of the body of scientific evidence for
36    effects  related to short- and long-term PMio-2.5 exposures.  In considering these causal inferences,
             Draft Do Not Quote or Cite            3-7                        March 2010

-------
 1    we take into account evidence of effects for which the reported associations provide suggestive
 2    evidence of a causal association. In considering the evidence, we have relied most heavily on the
 3    epidemiologic evidence, including our understanding of air quality distributions of PMi0-2.5 and
 4    PMio present during the times of the studies.  While being mindful of the inherent limitations and
 5    uncertainties in the currently available evidence, we have developed preliminary conclusions as
 6    to the adequacy of the current 24-hour PMio standard to protect against health effects associated
 7    with exposure to PMi0-2.5 and the degree to which alternative standards could be expected to
 8    protect against the reported health effects.
 9            In focusing on the key policy-relevant questions by which we have structured the
10    current review, our preliminary conclusions reflect upon our understanding of evidence-based
11    considerations to inform two overarching questions related to: (1) the adequacy of the current 24-
12    hour PMio standard to protect against effects associated with exposure to thoracic coarse particles
13    and (2) what potential alternative standard(s), if any, should be considered in this review.  In
14    addressing these broad questions, we have organized the discussions below around a series of
15    more specific questions reflecting different aspects of each overarching question.  When
16    evaluating the health protection afforded by the current or any alternative standards considered,
17    we have taken into account the four basic elements of the NAAQS (e.g., indicator, averaging
18    time, level, and form).
19           We believe that the approach outlined above, when presented in the final PA, will
20    provide a comprehensive basis to help inform the judgments required of the Administrator in
21    reaching decisions about the current and potential alternative primary standards meant to protect
22    public health against exposures to thoracic coarse particles.

23    3.2     ADEQUACY OF THE EXISTING PMio STANDARD
24           In considering the adequacy of the current 24-hour PMio standard to protect against
25    effects  associated  with exposures to thoracic coarse particles, staff addresses the following
26    overarching question:
27
28
29
30
Does the currently available scientific evidence, as reflected in the ISA, support or call into
question the appropriateness of maintaining a standard to protect against effects associated
 with exposure to thoracic coarse particles and the adequacy of the protection afforded by
                 the current 24-hour PMio standard against those effects?
31           To inform our consideration of this overarching question, we consider below the
32    evidence for a link between thoracic coarse particle exposures and adverse health effects (3.2.1),
33    including the evidence for the link between PMio-2.5 and mortality and morbidity; impacts on
34    susceptible populations; evidence for PMi0-2.5-related health effects in locations that meet the
             Draft Do Not Quote or Cite           3-8                         March 2010

-------
 1    current PMio standard; and uncertainties in the evidence. Preliminary staff conclusions
 2    regarding the adequacy of the current standard are presented in section 3.2.2. In considering the
 3    scientific information, we reflect upon both the information available in the last review and
 4    information that is newly available since the last review as assessed and presented in the ISA
 5    (US EPA, 2009a).

 6    3.2.1   Evidence of Effects Related to Ambient Thoracic Coarse Particles
 7           In the 2006 review of the PM NAAQS, the AQCD (US EPA,  2004, p. 9-48) concluded
 8    the following regarding thoracic coarse particles:
 9
10           For PMio-2.5, less evidence is available [than for PIVh.s or PMio], but the
11           studies using short-term exposures have reported results  that are of the
12           same magnitude as those for PMio and PIVh.s, though less often statistically
13           significant and thus having less strength, and the associations are generally
14           robust to alternative modeling strategies or consideration of potential
15           confounding by co-pollutants. This evidence is suggestive  of associations for
16           morbidity with short-term changes in PMio-2.5.
17
18           In contrast, the AQCD concluded that "[l]ittle evidence is available to allow conclusions
19    to be drawn about long-term PMio-2.5 exposures and morbidity" (US EPA, 2004, p. 9-46). In
20    considering these conclusions, the Administrator judged that "short-term exposure to thoracic
21    coarse particles can have an important public health impact" and that  available evidence
22    "suggests that there is a lack of such effects associated with long-term exposure to thoracic
23    coarse particles" (71 FR 61185/1, October 17, 2006). As noted above, the Administrator judged
24    that the then-existing body of scientific evidence supported retaining  a standard to protect against
25    health effects associated with short-term exposures to all thoracic coarse particles. Specifically,
26    the Administrator noted the following (71 FR 61185/1):
27
28           EPA continues to believe that the health evidence, including dosimetric, toxicologic, and
29           epidemiologic study findings, supports retaining a standard to protect against effects
30           associated with short-term exposure to thoracic coarse particles. As noted above and
31           summarized in section III. A of the proposal, there is a growing body of evidence
32           suggesting causal associations between short-term exposure to thoracic coarse particles
33           and morbidity effects, such as respiratory symptoms and hospital admissions for
34           respiratory diseases, and possibly mortality.
35
36           In considering the current body of scientific evidence for health effects of thoracic coarse
37    particles, we have considered the following question:
            Draft Do Not Quote or Cite           3-9                         March 2010

-------
 1    •  To what extent does the newly available scientific evidence and related uncertainties
 2       strengthen or call into question evidence of associations between ambient thoracic
 3       coarse particle exposures and health effects?
 4          We note that since the conclusion of the last review, the Agency has developed a more
 5    formal framework for reaching causal inferences from the body of scientific evidence, as
 6    discussed above in section 2.2.1.  Application of this framework draws upon the evaluation and
 7    synthesis of evidence from across epidemiologic, controlled human exposure, and toxicological
 8    studies.  This framework uses a five-level hierarchy that classifies the overall weight of evidence
 9    and causality using the following categorizations: causal relationship, likely to be a causal
10    relationship, suggestive of a causal relationship, inadequate to infer a causal relationship, and not
11    likely to be a causal relationship (ISA, section 1.5, Table 1-3).
12          Applying this framework to thoracic coarse particles, the ISA concludes that the existing
13    evidence is suggestive of a causal relationship between short-term PMio-2.5 exposures and
14    mortality, cardiovascular effects, and respiratory effects (US EPA, 2009a, section 2.3.3; see
15    Table 3-1 below). Several large multi-city epidemiologic studies, as well as a number of single-
16    city studies, have been conducted in the U.S. and Canada since the last review. These studies
17    have generally reported positive associations between ambient PMio-2.5 and morbidity and/or
18    mortality (US EPA, 2009a, section 2.3.4). The plausibility of associations reported in these
19    epidemiologic studies is supported by some experimental evidence, primarily from controlled
20    human exposure studies of heart rate variability and pulmonary inflammation, and by dosimetry
21    studies which show that a large proportion of inhaled particles in the 3-6 micron (dae) range can
22    reach and deposit in the lower respiratory tract, particularly the tracheobronchial  (TB) airways
23    (ISA, Figures 4-3 and 4-4).
24          However, important uncertainties remain with regard to the interpretation of this
25    evidence. For example, experimental  support for the associations reported in epidemiologic
26    studies has been  somewhat limited. Controlled human exposure studies have not reported effects
27    of thoracic coarse particles on pulmonary endpoints including lung function or respiratory
28    symptoms.  In addition, toxicological studies have not generally assessed inhalation of thoracic
29    coarse particles due to the technical challenges associated with conducting a PMio-2.5 inhalation
30    study in rodents. These studies have used intratracheal instillation and so provide only limited
31    support for the biological plausibility of the associations reported in epidemiologic studies (US
32    EPA, 2009a, section 2.3.4).  Beyond the limitations in experimental support, limitations in the
33    PMio-2.5 monitoring network, uncertainties in the ambient PMio-2.5 concentrations reported in
34    epidemiologic studies, the relatively small number of epidemiologic studies that have evaluated
35    co-pollutant models to address the potential for confounding by co-pollutants, variability in the

            Draft Do Not Quote or Cite          3-10                        March 2010

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
      chemical and biological composition of PMio-2.5, and limited evidence regarding effects of the
      various components of PMio-2.5, are also important sources of uncertainty (US EPA, 2009a,
      sections 2.3.3, 2.3.4).
            With respect to effects associated with long-term PMio-2.5 exposures, the ISA concludes
      that available evidence is inadequate to infer a causal relationship with all health outcomes
      evaluated (US EPA, 2009a, section 2.3). Specifically, similar to the judgment made in the
      AQCD (US EPA, 2004), the ISA states, "To date, a sufficient amount of evidence does not exist
      in order to draw conclusions regarding the health effects and outcomes associated with long-term
      exposure to PMi0-2.5" (US EPA, 2009a, section 2.3.4; see Table 3-1 below).

      Table 3-1.  Summary of Causality Determinations for PMi0-2.s
Exposure Duration
Short-term
Long-term
Outcome
Cardiovascular Effects
Respiratory Effects
Mortality
Central Nervous System Effects
Cardiovascular Effects
Respiratory Effects
Mortality
Reproductive and Developmental Effects
Cancer Mutagenicity, Genotoxicity Effects
Causal Determination
Suggestive
Suggestive
Suggestive
Inadequate
Inadequate
Inadequate
Inadequate
Inadequate
Inadequate
12
13
14
15
16
17
18
19
20
21
22
23
24
      Source: adapted from US EPA, 2009a; Table 2-6

      Therefore, in considering the health evidence more specifically below we have focused on
      mortality, cardiovascular effects, and respiratory effects associated with short-term exposures to
      thoracic coarse particles.
      Mortality
           We  note that, in the last review, a limited number of studies, mostly single-city analyses,
      were evaluated that examined thoracic coarse PM for its association with mortality (US EPA,
      2004).  Of those studies, a small number examined both PM2.5 and PMio-2.5 effects, and reported
      some evidence for PMio-2.5 effects of the same magnitude as PM2.5. Studies conducted in
      Phoenix (Mar et al., 2003) and Coachella Valley (Ostro et al., 2003) reported positive and
      statistically significant associations between PMio-2.5 and mortality while other single-city studies
      reported PMio-2.5 effect estimates that were positive, but not statistically significant.  Multiple
            Draft Do Not Quote or Cite
                                               3-11
March 2010

-------
 1    limitations were identified in these studies including measurement and exposure uncertainties for
 2    PMio-2.5 and the correlation between PM2.5 and PMio-2.5.  These limitations decreased the
 3    precision of effect estimates and increased the uncertainty surrounding the concentrations at
 4    which PMio-2.5-mortality associations were observed (US EPA, 2009a, section 6.5.2.3).
 5         Evidence that has become available since the last review provides additional support for
 6    the link between exposure to thoracic coarse particles and mortality, though important
 7    uncertainties remain regarding the extent to which PMi0-2.5 itself contributes to reported effects
 8    and the thoracic coarse particle concentrations at which associations occur (discussed below in
 9    more detail). The ISA assesses several recent studies that have evaluated associations between
10    mortality and PMio-2.5, most of which have reported positive, and in some cases, statistically
11    significant PMio-2.5 effect estimates (US EPA, 2009a, Figure 6-30).  This includes a recent U.S.-
12    based multicity study (Zanobetti and Schwartz, 2009) and a  Canadian multi-city study (Burnett
13    et al., 2004), both of which reported positive associations (though in the case of the Canadian
14    study, not statistically significant) between PMio-2.5 and mortality. In the U.S. study, a
15    significant association with PMio-2.5  was reported for all-cause, cardiovascular, and respiratory
16    mortality (US EPA, 2009a, section 6.5.2.3). The effect estimate for all-cause mortality remained
17    "relatively robust" (US EPA, 2009a, p. 6-82) and statistically significant in a two-pollutant
18    model that included PM2.5, while the effect estimates for  cause-specific mortality remained
19    positive but not statistically significant. When examining the city-specific effect estimates for
20    the 47 cities included in the thoracic coarse particle analysis, statistically significant increases in
21    mortality were associated with PMio-2.5 in St. Louis, MO; Salt Lake City, UT; Chicago, IL;
22    Pittsburgh,  PA; Detroit, MI; and Birmingham, AL. Positive, but not statistically significant,
23    associations were reported for all cause and/or cause-specific mortality in the remaining cities
24    (US EPA, 2009a, Figure 6-29). A number of additional studies have also reported positive, but
25    not significant, associations between PMio-2.5 and mortality (US EPA, 2009a, Figure 6-30).
26    Considered as a whole, the ISA notes that epidemiologic studies that have evaluated thoracic
27    coarse particles have reported consistent, positive associations between PMio-2.5 and mortality
28    (US EPA, 2009a, section 6.5.2.3).
29    Cardiovascular effects
30           With regard to cardiovascular morbidity, we note that the evidence evaluating
31    associations with short-term concentrations  of PMio-2.5 was limited in the last review.  Single-city
32    epidemiologic studies found generally positive associations, with some reaching statistical
33    significance, between PMio-2.5 and cardiovascular-related hospital admissions in Toronto
34    (Burnett et  al., 1997; 1999) and Detroit, MI (Ito, 2003). In the Detroit study, the PMio-2.5 effect
35    estimates for ischemic heart disease  (MD) remained positive and statistically significant in two-
             Draft Do Not Quote or Cite           3-12                        March 2010

-------
 1    pollutant models that included gaseous co-pollutants (US EPA, 2009a, Figure 6-5), while effect
 2    estimates remained positive and relatively unchanged in magnitude, but not statistically
 3    significant, for congestive heart failure (CHF). Effect estimates in the Toronto study (Burnett et
 4    al., 1997) were decreased and became non-significant in two-pollutant models that included
 5    gaseous co-pollutants (US EPA, 2009a, ISA, Figure 6-5). In addition, one study considered in
 6    the last review reported a positive, but not significant, association between onset of myocardial
 7    infarction (MI) and short-term PMi0-2.5 concentrations in Boston (Peters et al., 2001) and,
 8    although not a study of PMio-2.5 specifically, Schwartz et al. (1997) reported a statistically
 9    significant association between PMio and increased hospitalizations for cardiovascular disease in
10    Tucson, AZ, an urban area where thoracic coarse particles comprise a much greater fraction of
11    PMio than fine particles. No controlled human exposure or animal toxicological studies of PMi0.
12    2.5 and cardiovascular endpoints were presented in the 2004 AQCD (US EPA, 2004).
13           In addition to these cardiovascular morbidity studies considered in the last review, the
14    ISA assesses a recent multi-city study evaluating hospital admissions and emergency department
15    visits for cardiovascular disease in Medicare patients (Peng et al., 2008).  In this study of older
16    adults, the authors reported a positive and statistically significant association between 24-hour
17    PMio-2.5 concentrations and cardiovascular disease hospitalizations in a single pollutant model
18    using air quality data for 108 U.S. counties with one or more co-located PM2.5 and PMio
19    monitors. The effect estimate was reduced only slightly, though it was no longer statistically
20    significant, in two-pollutant models that included PM2.s (US EPA, 2009a, sections 2.3.3,
21    6.2.10.9). In addition to this U.S. multi-city study, positive associations reported for short-term
22    PMio-2.5 and cardiovascular morbidity reached statistical significance in a multi-city study in
23    France (Host et al., 2007) and associations were positive, but often did not reach statistical
24    significance, in several other locations (US EPA, 2009a, Figures 6-1 to 6-3, 6-5).  In considering
25    the available epidemiologic evidence, the ISA concludes that single- and multi-city
26    epidemiologic studies generally report positive associations between short-term PMio-2.5
27    concentrations and hospital admissions or emergency department visits for cardiovascular causes
28    (US EPA, 2009a, section 2.3.3, 6.2.12.2).
29           The generally positive associations between PMio-2.5 and cardiovascular morbidity
30    reported in these studies are supported by several recent epidemiologic studies that have
31    examined dust storm events and reported increases in cardiovascular-related emergency
32    department visits and hospital admissions5; by studies reporting positive associations with
      5Such dust storm studies are also important because they provide evidence that cardiovascular effects are associated
      with exposures to particles of non-urban origin.  As discussed more fully below, this evidence has implications for
             Draft Do Not Quote or Cite           3-13                         March 2010

-------
 1    cardiovascular mortality; by reported associations with other cardiovascular health effects
 2    including supraventricular ectopy and changes in heart rate variability; and by a limited number
 3    of controlled human exposure studies that have reported alterations in heart rate variability
 4    following exposure to PMi0-2.5 (US EPA, 2009a, sections 2.3.3, 6.2.12.2).  As noted above, the
 5    few toxicological studies that examined the effect of PMio-2.5 on cardiovascular health effects
 6    used intratracheal instillation and, as a result, provide only limited evidence on the biological
 7    plausibility of PMi0-2.5 induced cardiovascular effects (US EPA, 2009a, sections 2.3.3, 6.2.12.2).
 8    Respiratory effects
 9           With respect to respiratory effects associated with short-term PMio-2.5 exposures, we first
10    note that, in the last review, epidemiologic studies reported generally positive associations
11    between PMio-2.5 and respiratory-related hospitalizations or emergency department visits (e.g.,
12    for pneumonia, chronic obstructive pulmonary disease, and respiratory diseases combined) (US
13    EPA, 2004).  Support for these associations came from a small number of studies that examined
14    respiratory-related mortality and respiratory symptoms. Experimental evidence for respiratory
15    effects of PMio-2.5 was limited to a few animal toxicology studies and no controlled human
16    exposure studies.
17           In the current review, the ISA notes that a number of recent epidemiologic studies have
18    reported consistent, positive associations between short-term ambient PMio-2.5 concentrations and
19    respiratory-related emergency department visits and hospital admissions (US EPA, 2009a,
20    section 2.3.3). In a French multi-city study (Host et al., 2008), 24-hour ambient PMio-2.5
21    concentrations were positively associated with respiratory-related hospital admissions among
22    children, with an effect estimate larger than that for PM2.5. A U.S. multi-city study of Medicare
23    patients (Peng et al., 2009) reported a positive, but not statistically significant, effect estimate for
24    respiratory-related hospital admissions. A number of additional, mostly single-city, studies have
25    also reported positive, and in some  cases statistically significant, PMio-2.5 effect estimates for
26    respiratory-related hospital admissions and emergency department visits (ISA, Figures 6-10 to 6-
27    15).  In epidemiologic studies of respiratory effects, the strongest relationships with PMio-2.5
28    were observed among children, with less consistent evidence for adults and older adults (i.e., age
29    65 and older) (US EPA, 2009a, section 2.3.3.1). A limited number of epidemiologic studies
30    have focused on specific respiratory morbidity outcomes and found no evidence of an
31    association with lower respiratory symptoms,  wheeze, and medication use (US EPA, 2009a,
32    sections 2.3.3.1 and 6.3.1.1). While controlled human exposure studies have not observed an
      decisions on the appropriateness of the current PM10 standard as well as on the choice of indicator for thoracic
      coarse particles.
             Draft Do Not Quote or Cite           3-14                        March 2010

-------
 1    effect on lung function or respiratory symptoms in healthy or asthmatic adults in response to
 2    short-term exposure to PMio-2.5, healthy volunteers have exhibited an increase in markers of
 3    pulmonary inflammation. Toxicological studies using inhalation exposures are still lacking, but
 4    pulmonary injury and inflammation has been observed in animals after intratracheal instillation
 5    exposure (US EPA, 2009a, section 6.3.5.3) and, in some cases, PMio-2.5 was found to be more
 6    potent than PM2.5.
 7           Some studies of PMio-2.5 and respiratory morbidity have investigated potential
 8    confounding by co-pollutants through the application of co-pollutant models (US EPA, 2009a,
 9    section 6.3.8.5, Figure 6-15). Several of these studies have reported positive and statistically
10    significant PMio-2.5 effect estimates in co-pollutant models that included gaseous pollutants (e.g.,
11    Lin et al., 2002; Yang et al., 2004; Chen et al., 2005; Lin et al., 2005).  Several other studies
12    report that PMio-2.5 effect estimates remain positive, though not always statistically significant, in
13    co-pollutant models that include gaseous pollutants or PM2.5 (US EPA, 2009a, Figure 6-15).
14    Summary
15           In considering the extent to which newly available scientific evidence strengthens or calls
16    into question evidence of associations identified in the last review between ambient thoracic
17    coarse particle concentrations and adverse health effects, we recognize that the available
18    epidemiologic and experimental evidence now includes several recent  multi-city epidemiologic
19    studies conducted in the U.S., Canada, and Europe which have reported associations of PMio-2.5
20    with cardiovascular and respiratory morbidity and mortality, several studies of dust storm events
21    reporting associations with particles of non-urban origin, as well as some controlled human
22    exposure studies reporting effects of PMio-2.5 exposure on heart rate variability and pulmonary
23    inflammation. While uncertainties identified in the last review remain (see below), these recent
24    studies provide important information not  available in that review on the link between PMio-2.5
25    and mortality and morbidity, and  they have broadened our understanding of this link with
26    particles from different types of sources and in a variety of locations.  Thus, our preliminary
27    conclusion is that there is additional support, beyond that available in the last review, for
28    associations between adverse health effects (mortality, morbidity) and  short-term exposures to
29    PMio-2.5 from a broad mix of sources and a variety of locations.
30           Having reached this preliminary conclusion, we then consider how the new evidence
31    informs our understanding of susceptible populations by asking the following question:
32    •  To  what extent does the newly available scientific evidence expand our understanding
33       of susceptible populations, including identification of new susceptible populations?
34           Our understanding of populations that are more susceptible to PM exposures is discussed
35    above in chapter 2 (section 2.2.1). This includes populations that have a greater likelihood of
             Draft Do Not Quote or Cite          3-15                        March 2010

-------
 1    experiencing health effects related to exposure to PM due to a variety of factors including, but
 2    not limited to, genetic or developmental factors, race, gender, life stage, lifestyle (e.g., smoking
 3    status and nutrition), preexisting disease, or population-level factors that can increase an
 4    individual's exposure to PM (e.g., socioeconomic status, reduced access to health care, low
 5    educational attainment, residential location).  Although these studies have primarily evaluated
 6    exposures to PM2.5 or PMio, the available evidence suggests that the identified factors may also
 7    enhance susceptibility to PMi0-2.5 (US EPA, 2009a, section 2.4.1).  Given this, the discussion in
 8    section 2.2.1 of this document, which summarizes the evidence from the ISA as well as staff
 9    considerations and conclusions on susceptible populations, will not be repeated here.  Rather,
10    with regard to PMio-2.5, we note the overall conclusion from the ISA that "the  epidemiologic,
11    controlled human exposure, and toxicological studies evaluated in this review provide evidence
12    for increased susceptibility for various populations, including children and older adults, people
13    with pre-existing cardiopulmonary diseases, and people with lower SES"  (US EPA, 2009a, p. 2-
14    24). As discussed in more detail in section 2.2.1, these populations, which are of particular
15    concern with regard to effects linked to PM exposures,  are similar to the at-risk populations
16    considered in the last review of the PM NAAQS.
17           We next consider how the available evidence informs our understanding of air quality
18    concentrations associated with mortality and morbidity by asking the following question:
19    •  To what extent does  the newly available scientific evidence report associations that
20       extend to air quality levels that are lower than had previously been observed or that are
21       observed in areas that would likely meet the current PMio standard?
22           As discussed in more detail in chapter 2, the ISA concluded that there  is no evidence to
23    support the existence of a discernable threshold below which PM-associated effects would not
24    occur (US EPA, 2009a, section 2.4.3). Therefore, when considering the level  of protection
25    provided by the current 24-hour PMio standard against  exposures to thoracic coarse particles, we
26    consider whether available evidence provides support for associations between PMi0-2.5 and
27    mortality or morbidity in locations with PMio concentrations that would be allowed by the
28    current 24-hour PMio standard. In this section,  in considering air quality concentrations at which
29    health effects have been reported, we have  focused on effects for which the evidence is
30    suggestive of a causal relationship, as described in the ISA. Where sufficient  air quality data are
31    available, we have evaluated air quality concentrations  using the form of the current 24-hour
32    PMio standard.6
       The current form of the 24-hour PM10 standard is one expected exceedance per year, averaged over 3 years.
            Draft Do Not Quote or Cite          3-16                         March 2010

-------
 1           In the last review of the PM NAAQS, the Staff Paper (US EPA, 2005; section 5.4.1)
 2    noted that statistically significant associations between thoracic coarse particles and mortality
 3    had been reported in studies conducted in areas that did not meet the PMi0 standard during the
 4    time periods of the studies, including Phoenix (Mar et al., 2000, 2003), Coachella Valley, CA
 5    (Ostro et al., 2000, 2003), and Steubenville (as part of the Harvard Six Cities study, Schwartz et
 6    al.,  1996; Klemm et al., 2003).  In contrast, the Staff Paper noted that "[i]n areas with lower
 7    PMio-2.5 concentrations, no statistically significant associations were reported with mortality,
 8    though many were positive but not statistically significant"  (US EPA, 2005, p.  5-49).
 9           The Staff Paper also noted that epidemiologic studies of PMio, in areas where PMio is
10    typically dominated by the coarse fraction, provided additional supportive evidence for
11    associations between coarse fraction particles and health effects in areas with concentrations
12    generally not meeting the PMio standard levels. These studies included reports of associations
13    with hospitalization for cardiovascular diseases in Tucson, AZ (Schwartz, 1997); hospitalization
14    for chronic obstructive pulmonary disease in Reno/Sparks, NV (Chen et al., 2000);  medical visits
15    for asthma or respiratory  diseases in Anchorage, AK (Gordian et al., 1996; Choudhury et al.,
16    1997); and significant associations with mortality, respiratory hospital admissions,  and
17    respiratory symptoms in the Utah Valley area (e.g., Pope et al., 1989; 1991; 1992).
18           We have re-examined the issue of the PMio concentrations at which associations between
19    PMio-2.5 and mortality and morbidity have been reported in light of currently available evidence
20    and air quality information. In so doing, we have characterized PMio concentrations reported in
21    EPA's Air Quality System (AQS)7 from several U.S. study  locations (see US EPA, 2009a,
22    Figure 2-3). Specifically, consistent with the form of the current PMio standard, we consider the
23    second highest 24-hour average PMio concentration for each year of the  study,  from each
24    monitoring site in the study location. In an attempt to gain insight into whether the PMio air
25    quality concentrations in  a particular study location were above or below those allowed by the
26    current standard, we compare these second highest 24-hour PMio concentrations to the level of
27    the  standard, 150 ng/m3.8
28           PMio concentrations from U.S. study locations where positive and statistically significant
29    PMio-2.5 effect  estimates have been reported (e.g., US EPA,  2009a, Figures 2-3; 6-5; 6-24; 6-30)
30    are presented in Table 3-2 below for Detroit (Lippman, 2000; Ito et al., 2003; Zanobetti and
31    Schwartz, 2009), Phoenix (Mar et al., 2003), Seattle (Sheppard, 2003), Birmingham (Zanobetti
      7 http://www.epa.gov/ttn/airs/airsaqs/detaildata/downloadaqsdata.htm
      8We recognize that this is not the exact approach taken in making attainment and non-attainment designations for the
      PM10 standard, which also involves estimating expected exceedances for areas with less than daily monitoring, but it
      does provide important insights into the PM10 air quality in study areas, and into whether measured PM10 air quality
      concentrations would, or would not, be permitted under the current standard.
             Draft Do Not Quote or Cite           3-17                         March 2010

-------
1
2
3
4
5
6
7
      and Schwartz, 2009), Coachella Valley (Ostro et al., 2003), Chicago (Zanobetti and Schwartz,
      2009), Pittsburgh (Zanobetti and Schwartz, 2009), Salt Lake City (Zanobetti and Schwartz,
      2009), and St. Louis (Zanobetti and Schwartz, 2009).9
      Table 3-2. PMio Concentrations in Cities with Statistically Significant PMi0-2.s Effect
      Estimates
City

Birmingham
Coachella
Valley
Chicago
Detroit
(Zanobetti and
Schwartz)
Detroit
(Lippman;
Ito)10
Pittsburgh
Phoenix
Salt Lake City
Seattle
St. Louis
Second highest 24-hour PM10 concentration for each study location and study year
(Hg/m3)
Year
1
152
576
120
126
107
120
160
113
134
91
Year
2
157
278
123
113
113
123
130
117
138
92
Year
3
130
189
124
114
129
133
301
156
140
186
Year
4
160
108
106
96

107

123
131
224
Year
5
178
91
103
157

145

209
114
161
Year
6
166
84
84
139

153

149
119
191
Year
7
179
133
88
87

142

123
83
196
Year
8

155








Year
9

157








Year
10

114








 9
10
11
12
13
14
            Of these cities for which positive and statistically significant PMio-2.5 effect estimates
     have been reported, PMio concentrations in Birmingham, Phoenix, and St. Louis were higher
     than allowed by the current PMio standard during most of the study periods and concentrations in
     Coachella Valley were higher than allowed during about half of the study period.  In contrast,
     ambient PMio concentrations were below those allowed by the current standard during the entire
      9We have evaluated air quality in study locations for which single-city effect estimates are presented and for which
      air quality data during the study period are available.  Some important U.S. multi-city studies (e.g., Peng et al.,
      2008) are not included in Table 3-2 because they did not present single-city effect estimates.  We also note that there
      are several key studies conducted in Canadian cities; however, we were unable to obtain air quality information for
      individual monitors in these study locations.
      10These studies in Detroit presented effect estimates for two separate time frames.  Because PM10 FRM/FEM data is
      not available prior to  1988, we have focused here on the latter time frame used in the studies (1992-1994).
             Draft Do Not Quote or Cite           3-18                          March 2010

-------
 1    study periods for the studies conducted in Chicago, Detroit, Seattle, and Pittsburgh and during
 2    most of the study period in Salt Lake City.11
 3           We also note that two of the studies included in Table 3-2 evaluated PMi0-2.5 effect
 4    estimates in co-pollutant models.  Specifically, as discussed above, the associations reported by
 5    Ito et al. (2003) for PMio-2.5 and pneumonia hospital admissions in Detroit remained after
 6    adjustment for  gaseous co-pollutants.  In addition, the ISA reported that the overall PMio-2.5
 7    effect estimate  for all-cause mortality in the multi-city study by Zanobetti and Schwartz (2009),
 8    which included Chicago, Detroit, Seattle, Pittsburgh, and Salt Lake  City, remained "relatively
 9    robust" (US EPA, 2009a, p. 6-82) and statistically significant in a two-pollutant model that
10    included PM2.5, though co-pollutant models were not reported for individual cities.  As indicated
11    in the ISA (Figure 6-15), several other studies (e.g., those conducted in Canada) have also
12    reported robust and statistically significant PMio-2.5 effect estimates  in co-pollutant models,
13    however, we were unable to obtain air quality data for these other study locations and, therefore,
14    as described above, they were not included in Table 3-2.
15           In addition to the statistically significant results in the cities noted above, several studies
16    have reported positive, but not statistically significant, PMio-2.5 effect estimates  in locations with
17    PMio concentrations below those allowed by the current standard.  Specifically, Zanobetti and
18    Schwartz (2009) reported positive PMio-2.5 effect estimates for mortality in cities with a range of
19    PMio-2.5 concentrations, even down to the lowest PMio-2.5 concentrations estimated in the study
20    (ISA, figure 6-29). Many of these cities, where positive associations with mortality were
21    reported, attained the current PMio standard during the time period of the study.12 In addition,
22    Fairley et al.  (2003) reported a positive, but not statistically significant, association between
23    PMio-2.5 and mortality in Santa Clara County, CA (ISA, Figure 6-30).  Though some 24-hour
24    PMio concentrations in this location did exceed 150 |J,g/m3 (i.e., authors report concentrations up
25    to 165 ng/m3),  an analysis of Santa Clara air quality during the study period reveals that the 2nd
26    highest 24-hour PMio concentration did not exceed 150 ng/m3 during  any of the study years
27    (range was 58-147 ng/m3).  Peters et al. (2001) reported a positive, but not statistically
28    significant, association between PMio-2.5 and hospital admissions for myocardial infarction in
      nWe considered the number of times per year that 24-hour PM10 concentrations exceeded 150 |J.g/m3. For the
      studies in Chicago, Detroit, Seattle, and Pittsburgh, the average number of days per year with 24-hour PM10
      concentrations above 150 |J.g/m3 did not exceed 1 for any 3-year period. For Salt Lake City, year 3 of the study had
      2 exceedances of 150 |ag/m3 and year 5 had 3 exceedances.  Therefore, the average number of exceedances over
      years 3-5 was greater than 1.  The average over other 3 year periods during the study did not exceed 1.
      12 See http://www.epa.gov/oaqps001/greenbk/pindex.html for information on attainment status of specific locations
      and http://www.epa.gov/airtrends/values.html for information on design values for PM.
             Draft Do Not Quote or Cite           3-19                          March 2010

-------
 1    Boston, a location where the 2nd highest 24-hour PMio concentration did not exceed 150
 2    during either of the study years (52 and 80 |j,g/m3).
 3           A series of studies conducted in Atlanta have reported mixed results, but none
 4    statistically significant, for PMio-2.5 and mortality and respiratory-related emergency department
 5    visits. Specifically, Klemm et al. (2004), Metzger et al. (2004), Peel et al. (2005), and Tolbert et
 6    al. (2007) reported both positive and negative PMi0-2.5 effect estimates for mortality and
 7    emergency department visits (ISA, Figure 2-3). The 2nd highest 24-hour PMio concentrations in
 8    a year over the range of years encompassed by these studies did not exceed 150 ng/rn3
 9    (maximum was 111  |j,g/m3).
10    Summary
11           When viewed as a whole, these U.S. studies provide evidence for associations between
12    ambient PMio-2.5 concentrations and increased mortality and morbidity in several locations with
13    ambient PMio concentrations below those allowed by the current PMio standard. Staffs
14    preliminary conclusions regarding the potential implications of these studies for a decision on the
15    adequacy of the current PMio standard are discussed below.
16           We next consider the important uncertainties in the available evidence by asking the
17    following question:
18    •  What are the important uncertainties associated with the currently available scientific
19       evidence that should be considered in evaluating the adequacy of the current PMio
20       standard? Have these uncertainties changed  or been reduced since the last review?
21           The majority of the health evidence  supporting the link between thoracic coarse particle
22    exposures and adverse cardiovascular and respiratory effects comes from epidemiologic studies.
23    Staff notes that, while  several new studies have become available since the last review, little
24    progress has been made in reducing some of the important uncertainties inherent in these studies.
25    These uncertainties, and their implications for interpreting the scientific evidence, are discussed
26    below in more detail.
27           The ISA (sections 2.3.3, 2.3.4) concludes that an important uncertainty in the PMio-2.5
28    epidemiologic literature is that associated with the air quality estimates used in these studies.
29    Specifically, the ISA concludes that there is greater error in estimating ambient concentrations of
30    PMio-2.5 than in estimates for PM2.5 and,  therefore, that such uncertainty is a particularly relevant
31    consideration when interpreting PMio-2.5 epidemiologic studies.  Contributing to this uncertainty
32    is the relatively limited spatial coverage provided by existing PMio-2.5 monitors (US EPA, 2009a,
33    sections 2.2.3, 2.3.3, 2.3.4).  Currently, a national network to monitor PMio-2.5 is not in place,
34    limiting the spatial area over which PMio-2.5 concentrations are measured.  Specifically, of the
            Draft Do Not Quote or Cite          3-20                        March 2010

-------
 1    3,225 counties in the U.S., only 40 (1%), incorporating less than 5% of the population, met
 2    completeness and co-location criteria during the years 2005-2007, allowing analysis of PMio-2.5
 3    air quality in the ISA (US EPA, 2009a, section 3.5.1.1). In addition, based on the limited
 4    available evidence, the ISA concluded that "there is greater spatial variability in PMio-2.5
 5    concentrations than PM2.5 concentrations, resulting in increased exposure error for the larger size
 6    fraction" (US EPA, 2009a, p. 2-8) and that available measurements do not provide sufficient
 7    information to adequately characterize the spatial distribution of PMio-2.5 concentrations (US
 8    EPA, 2009a, section 3.5.1.1). As noted in the ISA, these limitations in estimates of ambient
 9    PMio-2.5 concentrations "would tend to increase uncertainty and make it more difficult to detect
10    effects of PMio-2.5 in epidemiologic studies" (US EPA 2009a, p. 2-21).
11           Uncertainty also results from the different approaches taken to estimate PMio-2.5
12    concentrations in epidemiologic studies.  The ISA notes that ambient concentrations of PMio-2.5
13    are generally determined by the subtraction of PM2.5 from PMio measurements, with different
14    studies using different methods. For example, one important multi-city study (e.g., Peng et al.,
15    2008) estimated PMio-2.5 by taking the difference between collocated PMi0 and PM2.5 monitors
16    while another important multi-city study (e.g., Zanobetti and Schwartz, 2009) used the difference
17    between county average PMio and PM2.5 concentrations. A small number of studies have
18    directly measured PMio-2.5 concentrations with dichotomous samplers (e.g., Burnett et al., 2004;
19    Villeneuve et al., 2003; Klemm et al., 2004).  It is not clear how computed PMio-2.5
20    measurements, such as those used by Zanobetti and Schwartz, compare with the PMio-2.5
21    concentrations obtained in other studies either by  direct measurement with a dichotomous
22    sampler or by calculating the difference using co-located samplers (US EPA, 2009a, section
23    6.5.2.3).13 Given the use of these different approaches to estimating PMio-2.5 concentrations
24    across studies, and their inherent limitations, the distributions of thoracic coarse particle
25    concentrations over which reported health outcomes occur remain highly uncertain.
26           The ISA also notes that the potential for confounding by co-occurring pollutants has been
27    addressed in only  a relatively small number of PMio-2.5 epidemiologic studies, introducing
28    additional uncertainty into the interpretation of these studies (US EPA, 2009a, section 2.3.3). As
29    discussed above, most studies that have evaluated co-pollutant models have reported that PMio-
30    2.5 effect estimates remain positive, but often lose precision and become statistically  non-
31    significant (US EPA, 2009a, Figures 6-5, 6-9, 6-15). In the U.S. multi-city study by Zanobetti
32    and Schwartz (2009) effect estimates for all-cause mortality remained statistically significant in a
      13In addition, when the difference between PM2 5 and PM10 is calculated, the potential for differences among
      operational flow rates and temperatures for PM10 and PM2 5 monitors add to the potential for exposure
      misclassification.
            Draft Do Not Quote or Cite          3-21                         March 2010

-------
 1    two-pollutant model that included PM2.5 while effect estimates for cause-specific mortality
 2    remained positive but were not statistically significant. Effect estimates for PMio-2.5 were also
 3    positive, but not statistically significant, in co-pollutant models that included PM2 5 in the studies
 4    by Peng et al. (2008) (cardiovascular and respiratory hospital  admissions) and Chen et al. (2004)
 5    (respiratory hospital admissions). In the Canadian multi-city study by Burnett et al. (2004) the
 6    PMio-2.5 effect estimates for respiratory-related hospital admissions remained positive, but not
 7    statistically significant, in a co-pollutant model that also included NO2, though the PMi0-2.5 effect
 8    estimate remained statistically significant in co-pollutant models with other gaseous pollutants.
 9    In other studies (e.g., see Lin et al., 2002; Ito et al., 2003; Chen et al., 2005) PMi0-2.5  effect
10    estimates remained relatively unchanged, and in some cases statistically significant, in co-
11    pollutant models that included gaseous pollutants.
12    Summary
13           As discussed above, important uncertainties remain in the evidence for associations
14    between PMio-2.5 and mortality  and morbidity.  In considering these uncertainties, we reach the
15    preliminary conclusion that, although a number of studies have been conducted since the last
16    review, the important uncertainties present in that review remain.  The implications of these
17    uncertainties for decisions on the adequacy of the PMio standard are discussed below.

18    3.2.2   Preliminary Staff Conclusions on Adequacy of Current PMio Standard
19           Collectively, taking into consideration the responses to specific questions focusing on
20    different ways to inform a decision on the adequacy of the current 24-hour PMio standard , we
21    revisit the overarching question:
22       •   Does the currently available scientific evidence, as reflected in the ISA, support or
23           call into question the appropriateness of maintaining a standard to protect against
24           effects associated with exposure to thoracic coarse particles and the adequacy of the
25           protection afforded by the current 24-hour PMio standard against those effects?
26           As an initial matter, we note that the current PMio standard is meant to protect the public
27    health against effects associated with exposures to all PMio-2.5. As discussed above, this was
28    judged in the last review to be appropriate given the "growing body of evidence suggesting
29    causal associations between short-term exposure to thoracic coarse particles and morbidity
30    effects, such as respiratory symptoms and hospital admissions for respiratory diseases, and
31    possibly mortality" (71 FR 61185, October 17, 2006). In considering the currently available
32    scientific evidence on thoracic coarse particles, we conclude that newly available studies
33    conducted since the last review of the PM NAAQS strengthen this conclusion.  As discussed

             Draft Do Not Quote or Cite           3-22                        March 2010

-------
 1    above, the newly available epidemiologic and experimental evidence, combined with evidence
 2    available in the last review, includes the following:
 3       •   Several multi-city epidemiologic studies conducted in the U.S., Canada, and Europe
 4           which have reported associations of PMi0-2.5 with cardiovascular and respiratory
 5           morbidity and mortality
 6       •   Studies that have reported that PMio-2.5 effect estimates remain positive, and in some
 7           cases statistically significant, in co-pollutant models
 8       •   Studies that have reported positive associations between dust storm events and
 9           cardiovascular effects, broadening the available evidence for associations between
10           adverse health effects and particles from a variety of sources, including non-urban
11           sources
12       •   Controlled human exposure studies of heart rate variability and pulmonary inflammation.
13
14           While uncertainties identified in the last review remain, these recent studies support those
15    available in the last review and provide important information on the link between PMio-2.5 and
16    cardiovascular and respiratory effects, including mortality. In addition, several recent dust storm
17    studies provide evidence not available in the  last review regarding associations between particles
18    of non-urban origin and adverse health effects. Therefore, consistent with the previous reviews,
19    we judge that it is appropriate to maintain a standard to protect the public health against effects
20    associated with exposures to all thoracic coarse particles.
21           In drawing preliminary conclusions regarding the adequacy of the level of protection
22    afforded by the current PMio standard, we have considered the ambient PMio concentrations in
23    locations where epidemiologic studies of thoracic coarse particles have been conducted as well
24    as the uncertainties associated with the broader body of scientific evidence.  Specifically, we
25    note that several recent U.S. studies have reported generally positive, and in  some cases
26    statistically significant, PMio-2.5 effect estimates in locations with 24-hour PMio concentrations
27    below those permitted by the current standard. This includes single-city and multi-city studies,
28    and studies that have reported that PMio-2.5 effect estimates remain positive, and in  some cases
29    robust and statistically significant, in co-pollutant models.  As described above, the decision in
30    the last review to retain the existing PMio standard was based, in part, on the conclusion that the
31    strongest evidence for the link between PMio-2.5 and health effects came from studies conducted
32    in locations with PMio concentrations above  those allowed by the standard.  Given that recent
33    epidemiologic studies have reported statistically significant PMio-2.5 effect estimates in several
34    locations with PMio concentrations below those allowed by the standard, a similar conclusion
35    would not be supported by the currently available evidence. Therefore, to the extent that the
             Draft Do Not Quote or Cite           3-23                        March 2010

-------
 1    approach to considering the adequacy of the current standard in this review is similar to the
 2    approach used in the last review, it would be appropriate to conclude that the current 24-hour
 3    PMio standard does not provide adequate public health protection and that it should be revised in
 4    order to increase protection against effects associated with short-term exposures to thoracic
 5    coarse particles.  If this approach to considering the evidence were adopted, a conclusion that the
 6    current standard is not adequate could be supported by the following observations:
 7         •   A number of epidemiologic studies have reported positive, and in some cases
 8             statistically significant, PMi0-2.5 effect estimates for mortality and morbidity in a variety
 9             of locations, including in several U.S.  cities with 24-hour PMio concentrations below
10             those allowed by the current PMio standard.
11         •   Studies that have evaluated co-pollutant models have generally reported that PMi0-2.5
12             effect estimates remain robust and, in some cases, statistically significant when PM2.5
13             or gaseous co-pollutants are added to the model.
14         •   Uncertainties in the extent to which ambient PMio-2.5 concentrations used in
15             epidemiologic studies reflect population exposures tend to bias the results of those
16             studies toward the null hypothesis.  Therefore, given limitations in PMio-2.5 monitoring,
17             the generally positive PMio-2.5 effect estimates reported across epidemiologic study
18             locations, even in locations for which effect estimates were not significant, provide
19             evidence for associations between PMio-2.5 and mortality and morbidity.
20         •   Controlled human exposure studies and, to a lesser extent, animal toxicological studies
21             have reported cardiovascular and respiratory effects following exposures to thoracic
22             coarse particles, thereby providing some support for the biological plausibility of the
23             associations reported in epidemiologic studies.
24
25           However, as discussed above, a decision on the adequacy of the level of public health
26    protection provided by the current PMio standard will be a public health policy judgment in
27    which the Administrator weighs the available evidence and its inherent uncertainties. Thus,
28    depending on the emphasis placed on different aspects of the evidence and its associated
29    uncertainties, different conclusions could be supported.  Specifically, as discussed above, we
30    note that many of the important uncertainties from the last review remain. In considering the
31    evidence as well as these uncertainties, the ISA concluded that the evidence is "suggestive" of a
32    causal relationship between short-term PMio-2.5 exposures and mortality, cardiovascular effects,
33    and respiratory effects. These conclusions contrast with those for associations between PM2.5
34    and health effects, which were judged to be either "causal" or "likely causal" for mortality  and
35    cardiovascular and respiratory effects associated with short- and long-term PM2.5 exposures.
36    Among the specific uncertainties noted in the ISA (US EPA, 2009a, sections 2.3.3, 2.3.4) are the
37    following:
             Draft Do Not Quote or Cite           3-24                        March 2010

-------
 1         •   Limited monitoring for PMio-2.5 results in uncertainty in the ambient PMio-2.5
 2             concentrations at which effects reported in epidemiologic studies occur.
 3         •   The number of epidemiologic studies that have employed co-pollutant models to
 4             address the potential for confounding remains relatively limited. Therefore, the extent
 5             to which PMio-2.5 itself contributes to reported health effects, rather than one or more
 6             co-pollutants, remains uncertain.
 7         •   Only a limited number of experimental studies provide support for the associations
 8             reported in epidemiologic studies.
 9         •   The chemical and biological composition of PMio-2.5, and the effects associated with
10             the various components, remains uncertain.
11
12    To the extent a decision on the adequacy of the current PMio standard were to emphasize the
13    uncertainties that contributed to the ISA conclusion that the evidence is "suggestive" of a causal
14    relationship for PMio-2.5, rather than indicating a "likely causal" or "causal" relationship, it would
15    be reasonable to conclude that the available evidence does not provide a basis for reaching a
16    fundamentally different conclusion from the one reached in the previous review (i.e., to retain
17    the current 24-hour PMio standard). Therefore, our preliminary conclusion is that the available
18    evidence could support either revising the current PMio standard to increase public health
19    protection against exposures to thoracic coarse particles or retaining the current PMio standard,
20    depending on the emphasis placed on different aspects of the evidence and associated
21    uncertainties.

22    3.3    CONSIDERATION OF POTENTIAL ALTERNATIVE STANDARDS
23          Having reached the conclusion that, depending on the approach to considering the
24    available evidence, the adequacy of the current 24-hour PMio standard to protect against effects
25    associated with exposures to thoracic coarse particles could be called into question, staff
26    considers a second overarching question:
27
28
What alternative standard(s) to protect against exposures to PMi0-2.s could be supported by
            the currently available scientific evidence, as reflected in the ISA?
29           To inform the answer to this overarching question, we consider how the currently
30    available scientific evidence could inform decisions regarding the basic elements of the NAAQS:
31    indicator (section 3.3.1), averaging time (section 3.3.2), and form and level (section 3.3.3).  In
32    considering the appropriateness of potential alternative standards, we consider both the evidence
33    available in the last review and the evidence that is newly available since the last review, as
34    assessed and presented in the ISA (US EPA, 2009a).
             Draft Do Not Quote or Cite          3-25                        March 2010

-------
 1    3.3.1   Indicator
 2           As noted in section 3.2 of the ISA, unlike gaseous pollutants such as O3, SO2, and CO,
 3    "which are well-defined chemical entities, atmospheric PM varies in size, shape, and chemical
 4    composition" (US EPA, 2009a, p. 3-1).  The selection of PMio as an indicator of thoracic
 5    particles was originally based in large part on dosimetry (U.S. EPA, 1996a) with evidence
 6    suggesting that a large proportion of inhaled coarse particles in the 3-6 um (dae) range can reach
 7    and deposit in the lower respiratory tract, particularly the tracheobronchial airways, while the
 8    fraction deposited decreases as particle size increases (US EPA, 2009a, Figures 4-3 and 4-4).
 9    Currently, it is most common to use PMio as an indicator of thoracic particles and PMio-2.5 as an
10    indicator of the thoracic component of coarse particles, sometimes referred to as thoracic coarse
11    particles. As such, the majority  of available health evidence for coarse particles links health
12    effects toPMio-2.5 and/or PMio.  As discussed in section 3.1.1, in the review of thePMNAAQS
13    completed in 1997, and again in the review completed in 2006, EPA concluded that the purpose
14    of the PMio standard would be to provide protection against effects associated with PMio-2.5.
15    This section considers the issue of the appropriate indicator for a standard meant to protect
16    against exposures to thoracic coarse particles.  Considerations related to indicator in the review
17    completed in 2006 are discussed in  section 3.3.1.1; evidence-based considerations to inform our
18    understanding of indicator in the current review are discussed in section 3.3.1.2; and preliminary
19    staff conclusions regarding indicator are discussed in section 3.3.1.3.

20    3.3.2   Consideration of Indicator in the Review Completed in 2006
21           In the last review, the Administrator's decision on the indicator for a standard meant to
22    protect against exposures to thoracic coarse particles took into consideration the potential for
23    particles originating in different types of environments (i.e., urban versus non-urban), and with
24    different compositions, to possess different degrees of toxicity. At the time, most of the studies
25    supporting the link between exposure to thoracic coarse particles and adverse health effects had
26    been conducted in urban locations.  Epidemiologic studies had reported positive, and in some
27    cases statistically significant, associations between ambient concentrations of thoracic coarse
28    particles and adverse health effects  in studies conducted in cities both inside and outside the U.S.
29    In contrast, very little evidence  was available to suggest that thoracic coarse particles from non-
30    urban areas posed a threat to  human health and a few studies reported that exposure to crustal
31    material from non-urban locations did not result in the types of effects that had  been linked  with
32    exposures to urban particles.  Specifically, several toxicological studies  reported that volcanic
33    ash from Mt. St. Helens (an  example of natural crustal material of geologic origin]  caused
34    very little toxicity in animal or in vitro model systems (e.g., see US EPA, 2005, section 5.4.2.1

            Draft Do Not Quote or Cite          3-26                        March 2010

-------
 1    for discussion}. In addition, an epidemiologic study in Spokane, WA specifically assessed
 2    whether mortality was increased on dust-storm days using case-control analysis methods.
 3    The average PMio concentration on dust storm days was more than 200 ug/m3 higher than on
 4    control days; however, the authors reported no evidence of increased mortality on these specific
 5    days (Schwartz et al., 1999). In addition, studies conducted in several areas in the western U.S.
 6    reported that associations between PMio and mortality or morbidity remained unchanged or
 7    became larger and more precise when days indicative of wind-blown dust were excluded from
 8    the analyses (Pope et al., 1999; Schwartz, 1997; Chen et al., 2000; Hefflin et al., 1994).  In the
 9    last review, this group of studies was interpreted as suggesting that health effects associated
10    with thoracic coarse particles are not driven by the types of natural crustal materials that
11    would typically form a major fraction of coarse particles in non-urban or rural areas.14
12           In 2006, the Administrator noted that the apparent differences in toxicity of
13    particles from urban versus non-urban locations could result from the different contaminants
14    present in urban versus non-urban environments. Specifically, he stated the following
15
16           [T]he observed toxicity of coarse particles in urban and industrial areas comes from the
17           kind of coarse particles found in these  environments, for example direct emissions from
18           industrial sources or materials released to road dust from motor vehicles such as brake
19           and tire wear, as well as from the contamination of coarse particles that can occur.  This
20           contamination can come from both mobile and stationary sources.  In particular, specific
21           components, such as byproducts of incomplete combustion  (e.g. poly cyclic aromatic
22           hydrocarbons) most commonly emitted from motor vehicles and other sources in the
23           form of PM2.5, as well as metals and other contaminants emitted from other
24           anthropogenic sources, appear in higher levels in urban areas (EPA, 2004a, p. 8-344; 71
25           FR 2665).  Many of these contaminants in PMio-2.5 come originally from fine particles,
26           which may become attached in the atmosphere or be deposited and mixed into coarse
27           materials on the ground. Thus the greater the concentration  of PM2.5, with higher levels
28           typically found in urban areas, the greater the level of contamination of coarse particles
29           by fine particles. This contamination increases the potential health risk posed by those
30           coarse particles. For that reason, it is logical  to allow lower levels of coarse particles
31           when fine particle concentrations are high (71 FR 61196, October 17, 2006).
32
33           Given the available evidence and the above considerations,  the Administrator judged in
34    the last review that two primary objectives were appropriate to consider when making a decision
35    on the standard indicator.  Specifically, given the preponderance of evidence from studies
36    conducted in urban areas, the Administrator judged it appropriate to have the level of protection
37    reflect the varying degree of public health concern presented by the different ambient mixes of
      14Although Ostro (2003) found that rural coarse particles primarily of crustal origin were associated with mortality
      effects, as noted by both CAS AC and the Administrator (71 FR 61190).
            Draft Do Not Quote or Cite           3-27                        March 2010

-------
 1    thoracic coarse particulate matter by allowing lower ambient thoracic coarse particle
 2    concentrations in locations with higher fine particle concentrations (e.g., urban areas, where the
 3    evidence indicates the public health risks to be significant) and higher concentrations of ambient
 4    thoracic coarse particles in locations with lower fine particle concentrations (e.g., non-urban
 5    areas, where the evidence related to thoracic coarse particles is more limited).  In addition,
 6    though there was little evidence at the time to suggest that thoracic coarse particles in non-urban
 7    locations posed a threat to human health, the Administrator judged that, given uncertainties in the
 8    interpretations of some of the studies of non-urban particles,15 it was appropriate to take a
 9    cautious approach by setting a standard that provides some measure of protection against
10    exposures to all ambient mixes of thoracic coarse particles, regardless of source of origin or
11    composition (71 FR 61197/3).
12           The Administrator evaluated several different potential indicators by considering the
13    extent to which each would contribute to achieving these objectives.  Specifically, the
14    Administrator considered a PMio-2.5 indicator, either qualified so as to exclude some sources and
15    locations (see below) or unqualified with no such exclusions, and a PMi0 indicator, either
16    adjusted to account for the  PM2.5 component of PMio or unadjusted.  Each of these options, and
17    the Administrator's consideration of them in the last review, is discussed below.
18           In considering the extent to which a PMio-2.5 indicator would achieve the objectives for
19    public health protection described above, the Administrator evaluated both a qualified indicator,
20    which would have exempted specific sources (e.g., agricultural and mining sources) and/or
21    locations (e.g., rural areas), and an unqualified indicator. With regard to the qualified indicator,
22    the Administrator concluded that, although it would permit the level of protection to reflect the
23    varying degree of public health concern presented by the different ambient mixes of thoracic
24    coarse particles, it would clearly not meet the goal of providing some measure of protection
25    against all ambient mixes of thoracic coarse particles for the evident reason that it would have
26    explicitly excluded certain  types of sources and locations.  In contrast, with regard to the
27    unqualified indicator the Administrator noted that, though it would provide protection against
28    exposures to all ambient mixes of thoracic coarse particles, "if such an indicator were utilized as
29    part of a standard with a single unvarying level, it would not reflect the critical difference  in
30    evidence regarding the relative public health risks associated with urban and non-urban thoracic
31    coarse particles" (71 FR 61195).  That is, he concluded that if the level were selected to provide
32    appropriate protection against effects associated with exposure to the ambient mixes typical of
      15 The final decision in the review completed in 2006 noted that uncertainties in interpreting these and other studies
      relevant to the consideration of effects associated with non-urban particles (e.g., changing activity patterns on high
      dust days in epidemiologic studies) suggested that it remained appropriate to provide some measure of protection
      against all thoracic coarse particles, even those from non-urban areas.
             Draft Do Not Quote or Cite           3-28                         March 2010

-------
 1    urban or industrial areas, the standard could be judged to be more stringent than necessary to
 2    protect against effects associated with exposure to the ambient mixes in non-urban areas.
 3    Conversely, if a less stringent level were adopted on the grounds that there is less certainty that
 4    the ambient mix in non-urban areas poses a health risk, then the standard could be judged not to
 5    provide sufficient protection from the ambient mix found in urban or industrial areas. In both
 6    instances a PMi0-2.5 standard could be judged not to be requisite (i.e., "not lower or higher than is
 7    necessary"  Whitman, 531 U.S. at 476.) to protect the public health with an adequate margin of
 8    safety.  In considering a potential approach to addressing this issue, the Administrator also
 9    concluded that it would not be appropriate to set different PMi0-2.5 standard levels for urban
10    versus non-urban locations given the "lack of evidence to support establishing specific
11    quantitative distinctions in level based on variations in coarse particle composition and
12    differential toxicity" (71 FR 61195).  Given all of these considerations, the Administrator
13    concluded that it would not be appropriate to set a standard with a PMio-2.5 indicator at the time
14    of the last review.
15           In considering the extent to which a PMi0 indicator would achieve the objectives for
16    public health protection described above, the Administrator evaluated both an adjusted indicator,
17    to account for the PM2.5 component of PMio, and an unadjusted indicator.  For the adjusted
18    indicator, the mass of PM2.5 monitored in excess of the 24-hour standard for PM2.5 would have
19    been subtracted from monitored PMio mass in order to avoid the double regulation of PM2.5 in
20    the situations where this would have  had the most regulatory consequence. Under this option, on
21    days when the measured concentration of PMio exceeded the level of the standard and the
22    measured concentration of PM2.5 exceeded the level  of the 24-hour PM2.5 standard, the amount of
23    PM2.s in excess of the 24-hour PM2.5  standard would have been subtracted from the total PMio
24    mass (71 FR 61197).  This option was judged by the Administrator to not be appropriate for two
25    reasons. First, as noted above, the Administrator had determined that there should be less
26    allowable coarse paniculate matter as PM2.5 levels increase because these are the conditions
27    under which PMio-2.5 tends to become more contaminated and could become more harmful.
28    Furthermore, this indicator would have inappropriately relaxed the level of protection afforded
29    by the existing 24-hour PMio standard because it would have allowed higher total PMio levels on
30    days with high PM2.s levels.  This indicator was thus directionally backwards; it decreased
31    protection when conditions warranted increased protection. For both of these reasons, the
32    Administrator rejected the adjusted PMio indicator.
33           In his final decision in the review completed in 2006, the Administrator judged that the
34    continued use of the unadjusted and unqualified PMio indicator was appropriate for a standard
35    meant to protect against effects associated with exposures to PMio-2.5. In reaching this judgment,
             Draft Do Not Quote or Cite           3-29                        March 2010

-------
 1    the Administrator concluded that the PMio indicator appropriately maintains some protection
 2    from all ambient mixes of thoracic coarse particles and provides a level of protection that reflects
 3    the varying degree of public health concern presented by the different ambient mixes.
 4    Specifically, the Administrator noted that both PM2.5 and PMio-2.5 contribute to PMio mass.
 5    Therefore, he noted that that the PMio standard would be expected to allow lower concentrations
 6    of thoracic coarse particles in locations with higher concentrations of PM2.5 (i.e., urban areas)
 7    and higher concentrations of thoracic coarse particles in locations  with lower concentrations of
 8    PM2.5 (i.e, non-urban areas). As  discussed above, this was judged appropriate because the
 9    available evidence indicated the potential for significant public health risks  associated with
10    exposures to thoracic coarse particles in urban areas, which generally have higher fine particle
11    concentrations than rural areas, but provided much less certainty regarding risks in non-urban
12    areas.
13           Given the Administrator's conclusion that these elevated fine particle concentrations in
14    urban locations could result in greater fine particle contamination  of coarse particles resulting in
15    increased potential health risks posed by coarse particles in urban  areas, he noted the following:
16
17           To the extent that use of a PMio indicator would result in any reduction in PM2.5
18           concentrations in an area, this would reduce the potential health risk from coarse particles
19           in the area as well. There is no certainty that the contribution of PM2.5 to the health risk
20           associated with exposure to contaminated coarse particles would be appropriately
21           addressed through the fine particle standards alone.  Thus,  to the extent that the inclusion
22           of the PM2.5 fraction in the PMio indicator amounts to double regulation of PM2.5, its
23           inclusion is non-duplicative and reasonable (71 FR 61196).
24
25           In reaching these judgments on the standard indicator, the Administrator recognized that
26    the relationship between allowable thoracic coarse particle concentrations and fine particle
27    concentrations is qualitative.  That is, the varying coarse particle concentrations allowed under
28    the PMio standard do not precisely correspond to the likely  variations in toxicity of thoracic
29    coarse particles in different areas. However, the Administrator concluded that, while currently
30    available information does not allow a more precise adjustment for relative toxicity, the PMio
31    standard is expected to generally ensure that the coarse particle levels allowed will appropriately
32    be lower in urban areas and higher in non-urban areas.  Given all of the above considerations, the
33    Administrator judged in the last review that the unqualified and unadjusted PMio indicator was
34    appropriate for a standard meant to protect against exposures to PMio-2.5.16
      16As discussed above (section 3.1.2), this decision was upheld on review by the D.C. Circuit, which endorsed every
      aspect of the Administrator's rationale in reaching its decision.  559 F. 3d at 533-538.
             Draft Do Not Quote or Cite           3-30                         March 2010

-------
 1    3.3.3  Evidence-Based Consideration of Indicator in the Current Review
 2          In the current review, in considering potential alternative standards, we have considered
 3    the following question with regard to indicator:
 4    •   To what extent does the currently available information provide support for retaining
 5       or revising the current PMi0 indicator?
 6          As in the last review, a critical issue in selecting an appropriate indicator for a standard
 7    meant to protect against exposures to thoracic coarse particles remains the extent to which
 8    available evidence supports the appropriateness of a standard that allows lower concentrations of
 9    PMio-2.5 in urban areas than rural areas. In considering this issue, we note that the ISA assesses
10    several epidemiologic  studies of short-term exposures to thoracic coarse particles that have been
11    conducted since the last review (US EPA, 2009a, Chapter 6). Similar to the studies considered
12    in the previous review, as described above in more detail,  most of these more recent studies have
13    been conducted in urban locations in the U.S., Canada, and Europe.  However, a few recent
14    studies have examined the health impacts of dust storm events, where crustal material was
15    generally indicated using PMio  (US EPA, 2009a, section 6.2.10.1).  Specifically, a study of a
16    dust storm in the Gobi desert, which transported PM across the Pacific Ocean reaching the
17    western U.S. in the spring of 1998, reported no excess risk of cardiac or respiratory hospital
18    admissions in the population of British Columbia's Lower Fraser Valley despite hourly PMio
19    concentrations greater than 100 ug/m3 and daily average concentrations several times greater
20    than normal (Bennett et al., 2006).  In contrast, Middleton et al. (2008)  reported that dust storms
21    in Cyprus were associated with a 4.7% (95% CI: 0.7-9.0) and 10.4% (95% CI: -4.7 to 27.9)
22    increase in risk of hospitalization for all causes and cardiovascular disease, respectively. PMio
23    concentrations in this study were much higher than typically reported in non-dust storm  studies,
24    with hourly PMio concentrations during dust storms frequently approaching 300 to 400 ug/m3
25    and a maximum 24-hour average PMio concentration during the study almost  10-fold higher than
26    the level of the current PMio standard  (i.e., 1,371 ug/m3).  In addition, Chan et al. (2008) studied
27    the effects of Asian dust storms on cardiovascular hospital admissions in Taipei, Taiwan and
28    reported significant adverse effects during 39 Asian dust events. Daily PMio concentrations
29    exceeded 150 ug/m3 on several occasions during the study period and reached a maximum of
30    between 200 and 250 ug/m3  (see Figure  1 in Chan et al., (2008)). Bell et al. (2008) analyzed
31    these data independently and concluded that Asian dust storms were positively associated with
32    risk of hospitalization for ischemic heart disease. In addition, Yang et al.  (2009) reported that
33    hospitalizations for congestive heart failure were elevated during or immediately following 54
34    Asian dust storm events, though effect estimates were not statistically significant.  Maximum 24-

            Draft Do Not Quote or Cite           3-31                         March 2010

-------
 1    hour PMio concentrations were not reported but the mean 24-hour concentration due to dust
 2    storm events was 112 ng/m3.  In a dust storm study where PMi0-2.5 was specifically evaluated,
 3    Perez et al. (2008) tested the hypothesis that outbreaks of Saharan dust exacerbate the effects of
 4    PMio-2.5 on daily mortality in Spain.  During Saharan dust days, the PMio-2.5 effect estimate was
 5    larger than on non-dust days, and it became statistically significant whereas it was not
 6    statistically significant on non-dust days.  On Saharan dust days, mean PMio-2.5 concentrations
 7    were about 10% higher than on non-Saharan dust days.  Several animal toxicological studies
 8    have also reported that exposure to PM2.5 from dust storms increased pulmonary inflammation
 9    and blood pressure (Lei et al., 2004;  Chang et al., 2007; US EPA, 2009a, sections 6.2.5.3,
10    6.3.3.3).
11           In considering the implications of these studies for a decision on the indicator of a
12    standard meant to protect against exposures to thoracic coarse particles, we first consider the
13    following question:
14    •  To what extent is it appropriate to maintain an indicator that provides some measure of
15       protection against all thoracic coarse particles?
16           With regard to this question,  we note that, though most of the evidence for associations
17    with morbidity and mortality continues to come from studies conducted in urban areas,17
18    associations reported in a small number of recent dust storm studies suggest that caution is
19    warranted in drawing conclusions about the relative toxicity of thoracic coarse particles from
20    urban versus non-urban environments. These studies of dust storm events, discussed above, raise
21    credible concerns regarding the potential for non-urban particles to cause adverse health effects,
22    though the dust storm-related PMio concentrations tend to be higher than those in many locations
23    where recent U.S. and Canadian studies have reported associations with PMio-2.5- At a minimum,
24    these recent dust storm studies indicate that exposure to relatively high concentrations of non-
25    urban particles is associated with health effects that are similar to those that, in other studies,
26    have been associated with urban particles.
27           Given the above considerations, as an initial matter we reach the preliminary conclusion
28    that, for a standard meant to protect against exposures to thoracic coarse particles, it remains
29    appropriate to provide some measure of protection against exposures to all ambient mixes of
30    thoracic coarse particles, regardless of their source of origin or composition.  In drawing this
31    preliminary conclusion, we recognize the results of epidemiologic studies assessed in the ISA,
32    which have reported positive, and in some cases statistically significant, effect estimates for
      1'Though in some cases study locations extended into areas surrounding urban centers and likely included some
      non-urban (e.g., suburban) areas.
             Draft Do Not Quote or Cite          3-32                        March 2010

-------
 1    thoracic coarse particles in a variety of locations, including studies of dust storm events where
 2    high particle concentrations were due to windblown crustal material.  We also recognize that the
 3    causality judgments for thoracic coarse particles in the ISA were made for "PMi0-2.5 as a whole
 4    regardless of origin, since PMi0-2.5-related effects have been demonstrated for a number of
 5    different environments" (US EPA, 2009a, p. 2-19).  Given these considerations, we note that
 6    either a PMio or a PMi0-2.5 indicator would be expected to provide protection against all ambient
 7    mixes of thoracic coarse particles, as long as these indicators were not qualified so as to exclude
 8    certain types of sources or locations.
 9           In addition, we have considered the following question with regard to indicator:
10    •  To what extent does it remain appropriate to draw distinctions between thoracic coarse
11       particles from urban and non-urban environments when making a decision on the
12       indicator for a standard meant to protect against exposures to those particles?
13          In considering this question, we note that recent studies do provide some evidence for the
14    toxicity  of particles from a variety of environments, including particles of non-urban origin.
15    Given these studies, it could be judged reasonable to adopt an indicator that does not allow
16    different concentrations of thoracic coarse particles in urban and rural areas.  As discussed
17    above, a PMi0-2.5 indicator would accomplish this.  In contrast, given the generally higher fine
18    particle  concentrations in urban areas compared to rural areas, a PMi0 indicator would be
19    expected to allow lower PMio-2.5 concentrations in urban areas than in rural areas.  Therefore,  to
20    the extent that it is judged appropriate to set a standard that does not allow different
21    concentrations of thoracic coarse particles in different types of environments, it would be
22    appropriate to consider a PMi0-2.5 indicator.
23          However, as discussed above, we also note that most of the evidence for positive
24    associations between  PMio-2.5 and morbidity and mortality, particularly evidence for these
25    associations at relatively low concentrations of PMio-2.5, continues to come from studies
26    conducted in locations where the PMio-2.5 is expected to be largely of urban origin. While recent
27    studies of dust storm  events have provided some additional information on the health effects
28    associated with particles of non-urban origin, the PMio concentrations due to dust storm events
29    in these  studies are generally higher, and in at least some cases considerably higher, than those in
30    the U.S. and Canadian cities where positive and statistically significant associations have been
31    reported in recent studies.  To the extent that these considerations are emphasized in a decision
32    on the indicator for a  standard, a reasonable conclusion would be that recent studies have not
33    fundamentally changed our understanding of the relative toxicity of urban versus non-urban
34    particles, and that available evidence continues to support the conclusion that thoracic coarse

            Draft Do  Not  Quote or Cite           3-33                         March 2010

-------
 1    particles of urban origin, where fine particle concentrations tend to be highest, are of particular
 2    concern.
 3           Given this conclusion, it would be reasonable to consider an indicator that focuses control
 4    on areas with ambient mixes known with greater certainty to be associated with adverse health
 5    effects (i.e., mixes generally present in urban areas) since such an indicator could provide public
 6    health benefits with the greatest degree of certainty. With regard to this, as discussed in more
 7    detail above, we note that the varying levels of thoracic coarse particles allowed by a PMi0
 8    indicator would be  expected to target protection to those locations (i.e., urban or industrial areas)
 9    where the strongest evidence has been observed for associations between adverse health effects
10    and exposures to thoracic coarse particles. Therefore, under this approach to considering the
11    evidence, a reasonable conclusion would be that a PMi0 indicator remains appropriate for a
12    standard meant to protect against exposures to thoracic coarse particles. Thus, as with a decision
13    on the adequacy of the current standard, different decisions on indicator could be judged to be
14    appropriate, depending  on the emphasis placed on different aspects of the  evidence and its
15    associated uncertainties.

16    3.3.4   Averaging  Time
17           In considering the appropriate averaging time(s) for the PMio standard in the last review,
18    the Administrator considered several epidemiologic studies that had reported statistically
19    significant associations between short-term (24-hour) exposure to PMi0-2.5 and various morbidity
20    effects as well as mortality.  Based primarily on these studies, he concluded that the available
21    evidence continued to support a 24-hour averaging time for a standard intended to control
22    thoracic coarse particles.  In contrast, given the relative lack of studies supporting a link between
23    long-term exposures to  thoracic coarse particles and morbidity or mortality, and given the
24    conclusion reached in the AQCD that the available evidence did not suggest a link between
25    morbidity or mortality and long-term exposure to PMio-2.5 (US EPA, 2004a, p. 9-79), the
26    Administrator further concluded that an annual coarse particle standard was not warranted at that
27    time. In reaching this decision, the Administrator also noted that a 24-hour standard would be
28    expected to provide protection against any as yet unidentified potential effects of long-term
29    exposure at ambient concentrations. Thus, in the 2006 review the Administrator retained the 24-
30    hour PMio standard and revoked the annual PMio standard (71 FR 61198-61199).
31           In the current review, we reconsider the extent to which the available evidence continues
32    to support these decisions by considering the following question:
33    •  To what extent does the available evidence continue to support a 24-hour averaging
34       time for  a standard meant to protect against effects associated with short-term
35       exposures to PMio-i.s?
             Draft Do Not Quote  or Cite           3-34                        March 2010

-------
 1          With regard to this question, we note the conclusions from the ISA regarding the weight
 2    of evidence for different averaging times as well as the studies on which those conclusions are
 3    based. Specifically, as discussed above (see Table 3-1 above), the ISA concludes that the
 4    existing  evidence is suggestive of a causal relationship between short-term PMio-2.5 exposures
 5    and mortality, cardiovascular effects, and respiratory effects (ISA, section 2.3.3).  This
 6    conclusion is based largely on epidemiologic studies which have primarily evaluated
 7    associations between 24-hour PMi0-2.5 concentrations and morbidity and mortality (e.g., see ISA,
 8    Figure 2-3), though a smaller number of controlled human exposure studies have reported effects
 9    following shorter exposures (i.e., 2-hours) to PMio-2.5 (e.g., see ISA, sections 6.2.1.2, 6.3.3.2).  In
10    contrast, with respect to long-term exposures, the ISA concludes that available evidence is
11    inadequate to infer a causal relationship with all health outcomes evaluated (US EPA, 2009a,
12    section 2.3). Specifically, the ISA states, "To date, a sufficient amount of evidence does not
13    exist in order to draw conclusions regarding the health effects and outcomes associated with
14    long-term exposure to PMi0-2.5" (US EPA, 2009a, section 2.3.4;  see Table 3-1 below).
15          In considering the weight of evidence determinations in the ISA, we conclude that, at a
16    minimum, they suggest the importance of having a standard that protects against short-term
17    exposures to thoracic coarse particles.  In considering the averaging times used in the short-term
18    epidemiologic studies, we note that the majority of the evidence supporting the link between
19    PMio-2.5 and morbidity and mortality is based on 24-hour average thoracic coarse particle
20    concentrations.  Therefore, our preliminary conclusion is that the evidence available in this
21    review continues to support the appropriateness of a 24-hour averaging time for a PMio standard
22    meant to protect against short-term exposures to PMio-2.5, including potential effects associated
23    with exposures of shorter duration than 24-hours.  We further note that,  given the ISA conclusion
24    that "a sufficient amount of evidence does not exist in order to draw conclusions regarding the
25    health effects and outcomes associated with long-term exposure to PMio^.s" (ISA, p. 2-19) the
26    evidence does not support the appropriateness of an annual thoracic coarse particle standard at
27    this time. In reaching this conclusion, we note that, to the extent a short-term standard requires
28    areas to reduce their 24-hour ambient particle concentrations, long-term concentrations would
29    also be expected to decrease. Therefore,  a 24-hour standard meant to protect against short-term
30    exposures to thoracic coarse particles would also be expected to provide protection against any
31    as yet unidentified potential effects of long-term exposures at ambient concentrations.

32    3.3.5  Level and Form
33          Given the conclusions above regarding the adequacy of the current PMio standard (i.e.,
34    that, depending on the approach to considering the available evidence, the adequacy of the

            Draft Do Not Quote or Cite          3-35                        March 2010

-------
 1    current 24-hour PMio standard to protect against effects associated with exposures to thoracic
 2    coarse particles could be called into question) and the appropriate indicator (i.e., that, depending
 3    on how the evidence is considered, either a PMio or PMi0-2.5 indicator could be judged
 4    appropriate), we next consider whether different standard levels and/or forms for a 24-hour PMio
 5    or PMio-2.5 standard could be supported by the available evidence. To inform our consideration
 6    of this issue, we have considered the following question:
 7    •  To what extent does new information support consideration of an alternative form
 8       and/or level?
 9       In considering this question, we have taken into account the scientific evidence from studies
10    of the link between short-term exposure to PMio-2.5 and mortality, cardiovascular morbidity, and
11    respiratory morbidity, as assessed in the ISA, including the uncertainties and limitations in that
12    evidence.
13           In  the 2006 review, the Administrator concluded that "the level of protection afforded by
14    the current 24-hour PMio standard of 150 ug/m3, one-expected-exceedance form, continues to be
15    appropriate for the types of thoracic coarse particles typically found in urban or industrial areas"
16    (71 FR 61202).  In support of this decision, the Administrator noted that "mortality effects
17    observed in  epidemiologic studies for [thoracic] coarse particles are generally associated with
18    exposure levels that exceed the current standards, and morbidity effects are generally associated
19    with exposure levels that exceeded the current standards on only a few occasions" (71 FR
20    61202). Given this evidence, the Administrator concluded that the level of protection afforded
21    by the existing PMio standard was not greater than warranted.  In addition, the Administrator
22    concluded that uncertainties in population exposures in available morbidity studies18 suggested
23    that there was little basis for concluding that a greater degree of protection was warranted.
24    Therefore, the Administrator retained the existing level of 150 ug/m3 for the 24-hour PMio
25    standard, and the existing one-expected exceedance form.
26           In  the sections below, we discuss the approach that will be used to identify potential
27    alternative forms and a range of potential alternative levels for a standard meant to protect
28    against exposures to thoracic  coarse particles. These potential alternative standard forms and
29    levels will be presented and considered in the second draft PA document.
      18Analysis of air quality information in Detroit (Ito, 2003) and Coachella Valley (Ostro et al, 2003) suggested that
      population exposure concentrations in these areas could be appreciably higher (Detroit) or lower (Coachella Valley)
      than indicated by the air quality information used in the studies (See US EPA, 2005, pp. 5-64 to 5-66).
             Draft Do Not Quote or Cite           3-36                         March 2010

-------
 1    3.3.6   Form
 2           As discussed in more detail above, in the 2006 review the Administrator concluded that
 3    "the degree of protection afforded by the current 24-hour PMio standard is requisite to protect
 4    public health with an adequate margin of safety" (71 FR 61201).  Given the overall decision in
 5    that review to provide the same protection as provided by the existing 24-hour standard, the
 6    Administrator concluded "it is best to retain both the form and the level of the current primary
 7    24-hour PMio standard" (71 FR 61202). Therefore, the one-expected-exceedance form was
 8    retained.
 9           This decision is being examined in the current review within the context of the overall
10    decision on whether to retain or revise the current 24-hour PMio standard. While the selection of
11    a specific form must be made within the context of a decision on the other elements of the
12    standard, EPA generally favors the concentration-based form for short-term standards.  In 1997
13    EPA established 98th percentile form for the 24-hour PM2.5 standard and, most recently, in 2010
14    EPA established 98th percentile form for the 1-hour NO2 standard (62 FR 38671; 75 FR 6474).
15    In making these decisions, EPA judged that, as compared to an exceedance-based form, such as
16    that used for the current PMio standard, a concentration-based form is more reflective of the
17    health risks posed by elevated pollutant concentrations because it gives proportionally greater
18    weight to days when concentrations are well above the level of the standard than to days when
19    the concentrations are just above the standard.  Further, a concentration-based form better
20    compensates for missing data and less than-daily monitoring. In addition, when averaged over 3
21    years, it has greater stability than  an exceedance-based form and, therefore, it facilitates the
22    development of more stable implementation programs.  In these previous reviews, after
23    considering different concentration percentiles (95th to the 99th for PM2.5 and 98th and 99th for
24    NO2), EPA selected the 98th percentile as an appropriate balance between adequately limiting
25    the occurrence of peak concentrations and providing increased stability and robustness.
26           In considering these decisions, our preliminary conclusion is that a similar logic applies
27    in the case of the 24-hour PMio standard. For this reason, to  the extent that it is judged
28    appropriate to revise the current 24-hour standard  in the current review, we will give  preliminary
29    consideration to a range of alternative standard options based on 98th percentile concentrations of
30    PMio and/or PMio-2.5, as discussed below.
31    3.3.7   Level
32           As noted above, depending on the approach taken in this review to consider the evidence
33    for the adequacy of the current PMio standard, it could be judged that this standard does not
34    provide adequate public health protection. If this is judged to be the case, we will consider

             Draft Do Not Quote or Cite           3-37                        March 2010

-------
 1    potential alternative standard levels for a 24-hour standard that, in conjunction with a 98th
 2    percentile form as described above, would provide increased public health protection against
 3    effects associated with exposures to PMio-2.5. Given the conclusion above that the continued use
 4    of the PMio indicator could be judged appropriate for a standard meant to protect against
 5    exposures to PMio-2.5, we will consider a range of potential alternative standard levels for a PMio
 6    standard. However, as is also discussed above, a PMio-2.5 indicator could potentially be judged
 7    appropriate.  While the air quality information available to inform the identification of a range of
 8    potential alternative standard levels will be more limited for a PMio-2.5 standard than for a PMio
 9    standard, we recognize that some such information exists.  Therefore, to the extent that feedback
10    received during the review of this first draft Policy Assessment justifies consideration of a PMio-
11    2.5 standard, we will  discuss potential alternative levels for such a standard in the second draft
12    document.
13          In considering potential alternative standard levels, staff will consider the available
14    information on PMio, and possibly PMio-2.5, concentrations in locations where health studies
15    have evaluated the link between short-term PMio-2.5 concentrations and mortality, cardiovascular
16    morbidity, and respiratory morbidity, as well as the uncertainties and limitations in that evidence.
17    In so doing, we will focus on U.S. studies, as described above in our consideration of the
18    adequacy of the current standard. Specifically, we will consider the extent to which these studies
19    report positive and relatively precise effect estimates; the extent to which the reported
20    associations are robust in co-pollutant models; and the extent to which the studies used relatively
21    reliable  air quality data.  In particular, we will focus on those specific studies, briefly
22    summarized above and presented in more detail in chapter 6 of the ISA, that provide evidence of
23    associations between PMio-2.5 and mortality and morbidity in areas that would likely have met
24    the current 24-hour PMio standard during the time of the study. Staff believes that this body of
25    evidence can serve as a basis for identifying and considering potential alternative 24-hour PMio
26    and/or PMio-2.5 standards that would provide increased protection against effects related to short-
27    term exposures to PMio-2.5.
28          As an initial  matter, we recognize, as discussed above, that these short-term exposure
29    studies provide no evidence of clear thresholds, or lowest-observed-effects levels,  in terms of 24-
30    hour average concentrations. As a consequence, this body of evidence is difficult to translate
31    directly  into a  specific 24-hour standard that would independently protect against all effects
32    associated with short-term exposures. In the absence of an apparent threshold, for purposes
33    of identifying the range of standard levels potentially supported by the epidemiologic
34    evidence, we will focus on the range of PMio and/or PMio-2.5 concentrations that have been
35    measured in locations where key health studies have reported associations with PMio-2.5.
36    Specifically, we will focus on the upper end of the distributions of daily PM concentrations,
            Draft Do Not Quote or Cite          3-38                         March 2010

-------
 1    particularly in terms of the 98th percentile of the yearly distribution of 24-hour PMio and/or

 2    PMio-2.5 concentrations, reflecting the considerations above for the appropriate form of the 24-
 3    hour standard.  As noted above, this analysis of potential alternative standard levels will be
 4    presented in the second draft PA document.


 5    3.4     KEY UNCERTAINTIES AND AREAS FOR FUTURE RESEARCH AND DATA
 6            COLLECTION

 7            [This topic will be discussed in the second draft Policy Assessment.]


 8    3.5     REFERENCES

 9    Bell ML; Ebisu K; Peng RD; Walker J; Samet JM; Zeger SL; Dominic F (2008). Seasonal and regional short-term
10            effects of fine particles on hospital admissions in 202 U.S. counties, 1999-2005. Am J Epidemiol, 168:
11            1301-1310.

12    Bennett CM; McKendry IG; Kelly S; Denike K; Koch T (2006). Impact of the 1998 Gobi dust event on hospital
13            admissions in the Lower Fraser Valley, British Columbia. Sci Total Environ, 366: 918-925.

14    Burnett RT; Cakmak S; Brook JR; Krewski D (1997). The role of paniculate size and chemistry in the association
15            between summertime ambient air pollution and hospitalization for cardiorespiratory diseases. Environ
16            HealthPerspect, 105: 614-620.

17    Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M;  Vincent R; Dann T (2004). Associations between
18            short-term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup Health, 59:
19            228-236.

20    Chan CC; Chuang KJ; Chen WJ; Chang WT; Lee CT;  Peng CM (2008). Increasing cardiopulmonary emergency
21            visits by long-range transported Asian dust storms in Taiwan. Environ Res, 106: 393-400.

22    Chang C-C; Hwang J-S; Chan C-C; Wang P-Y; Cheng T-J  (2007). Effects of concentrated ambient particles on
23            heart rate, blood pressure, and cardiac contractility in spontaneously hypertensive rats during a dust storm
24            event. Inhal Toxicol, 19: 973-978.

25    Chen L; Yang W; Jennison BL; Omaye ST (2000). Air paniculate pollution and hospital admissions for chronic
26            obstructive pulmonary  disease in Reno, Nevada. Inhal Toxicol, 12: 281-298.

27    Chen Y; Yang Q; Krewski D; Shi Y; Burnett RT; McGrail K (2004). Influence of relatively low level of paniculate
28            air pollution on hospitalization for COPD in elderly people. Inhal Toxicol, 16: 21-25.

29    Chen Y; Yang Q; Krewski D; Burnett RT; Shi Y; McGrail KM (2005). The effect of coarse ambient paniculate
30            matter on first, second, and overall hospital admissions for respiratory disease among the elderly. Inhal
31            Toxicol, 17: 649-655.

32    Choudhury, A. H.; Gordian, M.  E.; Morris, S. S. (1997) Associations between respiratory illness and PMIO air
33            pollution. Arch. Environ. Health 52: 113-117.

34    Fairley D (2003). Mortality and air pollution for Santa Clara County, California, 1989-1996, In: Revised analyses of
3 5            time-series studies of air pollution and health. Special report. Health Effects Institute. Boston,
36            MA.http://www.healtheffects.org/Pubs/TimeSeries.pdf.
              Draft Do Not Quote or Cite           3-39                          March 2010

-------
  1     Gordian ME; Ozkaynak H; Xue J; Morris SS; Spengler JD (1996). Paniculate air pollution and respiratory disease in
  2            Anchorage, Alaska. Environ Health Perspect, 104: 290-297.

  3     Hefflin BJ; Jalaludin B; McClure E; Cobb N; Johnson CA; Jecha L; Etzel RA (1994). Surveillance for dust storms
  4            and respiratory diseases in Washington State, 1991. Arch Environ Occup Health, 49: 170-174.

  5     Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V;
  6            Prouvost H; Lefranc A (2007). Short-term Associations between Fine and Coarse Particles and
  7            Cardiorespiratory Hospitalizations in Six French Cities. Occup Environ Med, 18: S107-S108.

  8     Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V;
  9            Prouvost H; Lefranc A (2008). Short-term associations between fine and coarse particles and hospital
10            admissions for cardiorespiratory diseases in six French cities. Occup Environ Med, 65: 544-551.

11     Ito K (2003). Associations of paniculate matter components with daily mortality and morbidity in Detroit,
12            Michigan, In: Revised analyses of time-series studies of air pollution and health. Special report. Health
13            Effects Institute.  Boston, MA. R828112. http: //www.healtheffects.org/Pubs/TimeSeries.pdf.

14     Klemm RJ; Mason R (2003). Replication of reanalysis of Harvard Six-City mortality study. In HEI Special Report:
15            Revised Analyses of Time-Series Studies of Air Pollution and Health, Part II (pp. 165-172). Boston, MA:
16            Health Effects Institute.

17     Klemm RJ; Lipfert FW; Wyzga RE; Gust C (2004). Daily mortality and  air pollution in Atlanta: two years of data
18            from ARIES. Inhal Toxicol, 16 Suppl 1: 131-141.

19     Lei Y-C; Chan C-C; Wang P-Y; Lee C-T; Cheng T-J (2004). Effects of Asian dust event particles on inflammation
20            markers in peripheral blood and bronchoalveolar lavage in pulmonary hypertensive rats. Environ Res,  95:
21            71-76.

22     Lei YC; Chen MC; Chan CC; Wang PY; Lee CT; Cheng TJ (2004). Effects of concentrated ambient particles on
23            airway responsiveness and pulmonary inflammation in pulmonary hypertensive rats. Inhal Toxicol, 16:
24            785-792.

25     Lin M; Chen Y; Burnett RT; Villeneuve PJ; Krewski D (2002). The influence of ambient coarse particulate matter
26            on asthma hospitalization  in children: case-crossover and time-series analyses. Environ Health Perspect,
27            110:575-581.

28     Lippmann M; Ito K; Nadas A; Burnett RT (2000). Association of particulate matter components with daily mortality
29            and morbidity in urban populations. Health Effects Institute. Cambridge, MA.

30     Mar TF; Norris GA; Koenig JQ; Larson TV (2000). Associations between air pollution and mortality in Phoenix,
31            1995-1997. Environ Health Perspect, 108: 347-353.

32     Mar, T. F.; Norris, G. A.; Larson, T. V.; Wilson, W. E.; Koenig, J. Q. (2003). Air pollution and cardiovascular
33            mortality in Phoenix, 1995-1997. In: Revised analyses of time-series studies of air pollution and health.
34            Special report. Boston, MA: Health Effects Institute; pp. 177-182. Available:
35            http://www.healtheffects.org/Pubs/TimeSeries.pdf. October 18, 2004.

36     Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin  H (2004).
3 7            Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15: 46-56.

3 8     Middleton N; Yiallouros P; Kleanthous S; Kolokotroni O; Schwartz J; Dockery DW; Demokritou P; Koutrakis  P
39            (2008). A 10-year time-series analysis of respiratory and cardiovascular morbidity in Nicosia, Cyprus: the
40            effect of short-term changes in air pollution and dust storms. Environ Health, 7:39.

              Draft Do Not Quote or Cite            3-40                            March 2010

-------
  1     Miller, F.J; Gardner, D.E.; Graham, J.A.; Lee, R.E.; Wilson, W.E.; Bachmann, J.D. (1979). Size considerations for
  2            establishing a standard for inhalable particles. J Air Pollution Control Assoc. 29:610-615.

  3     Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2000) Coarse and fine particles and daily mortality in the Coachella
  4            Valley, California: a follow-up study. J. Exposure Anal. Environ. Epidemiol. 10: 412-419.

  5     Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2003). Coarse particles and daily mortality in Coachella Valley,
  6            California. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston,
  7            MA: Health Effects Institute; pp. 199-204. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf.
  8            October 18, 2004.

  9     Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H (2005).
10            Ambient air pollution and respiratory emergency department visits. Epidemiology, 16: 164-174.

11     Peng RD; Chang HH;  Bell ML; McDermott A; Zeger SL; Samet JM; Dominici F. (2008). Coarse paniculate matter
12            air pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare patients.
13            JAMA, 299: 2172-2179.

14     Peng R; Bell M; Geyh A; McDermott A; Zeger S; Samet J; Dominici F (2009). Emergency admissions for
15            cardiovascular and respiratory diseases and the chemical composition of fine particle air pollution. Environ
16            HealthPerspect, 117: 957-963.

17     Peters A; Dockery DW; Muller JE; Mittleman MA (2001). Increased paniculate air pollution and the triggering of
18            myocardial infarction. Circulation, 103: 2810-2815.

19     Pope CA III (1989). Respiratory disease associated with community air pollution and a steel mill, Utah Valley. Am J
20            Public Health, 79: 623-628.

21     Pope CA III; Dockery DW; Spengler JD; Raizenne ME (1991). Respiratory health and PM10 pollution: a daily time
22            series analysis. Am J Respir Crit Care Med, 144: 668-674.

23     Pope, C. A., Ill; Schwartz, J.; Ransom, M. R. (1992) Daily mortality and PM10 pollution in Utah valley. Arch.
24            Environ. Health 47: 211-217.

25     Pope, C. A., Ill; Hill, R. W.; Villegas, G. M. (1999) Paniculate air pollution and daily mortality on Utah's Wasatch
26            Front. Environ.  Health Perspect.: 107: 567-573.

27     Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically with fine particles? J.  Air
28            Waste Manage. Assoc. 46: 927-939.

29     Schwartz J (1997). Air pollution and hospital admissions for cardiovascular disease in Tucson. Epidemiology, 8:
30            371-377.

31     Schwartz J; Norris G; Larson T;  Sheppard L; Claiborne C; Koenig J (1999). Episodes of high coarse particle
32            concentrations are not associated with increased mortality. Environ Health Perspect, 107: 339-342.

33     Sheppard L; Levy D; Norris G; Larson TV; Koenig JQ (2003). Effects of ambient air pollution and nonelderly
34            asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology, 10: 23-30.

3 5     Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA (2007). Multipollutant modeling issues in a study of ambient air
36            quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 17: S29-S35.

37     US EPA (2004). Air Quality Criteria for Paniculate Matter. National Center for Environmental Assessment, Office
38            of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711;

              Draft Do Not Quote or Cite             3-41                            March 2010

-------
  1            report no. EPA/600/P-99/002aF and EPA/600/P-99/002bF. October 2004.  Available:
  2            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_cr_cd.html

  3    US EPA (2005). Review of the National Ambient Air Quality Standards for Paniculate Matter: Policy Assessment
  4            of Scientific and Technical Information, OAQPS Staff Paper. Research Triangle Park, NC 27711: Office
  5            of Air Quality Planning and Standards; report no. EPA EPA-452/R-05-005a. December 2005. Available:
  6            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_cr_sp.html

  7    US EPA (2008a).  Integrated Review Plan for the National Ambient Air Quality Standards for Paniculate Matter.
  8            National  Center for Environmental Assessment and Office of Air Quality Planning and Standards, U.S.
  9            Environmental Protection Agency, Research Triangle Park, NC. Report No. EPA 452/R-08-004. March
10            2008. Available at:
11            http://www.epa.gov/ttn/naaqs/standards/pm/data/2008_03_final_integrated_review_plan.pdf

12    US EPA (2009a).  Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
13            Protection Agency, Washington, DC, EPA/600/R-08/139F, December 2009. Available:
14            http://cfpub.epa.gov/ncea/cfm/recordisplay .cfm?deid=216546

15    U.S. EPA. (2009b). Paniculate Matter National Ambient Air Quality Standards: Scope and Methods Plan for Health
16            Risk and Exposure Assessment. U.S. Environmental Protection Agency, Research Triangle Park, NC,
17            EPA-452/P-09-002.  February 2009. Available:
18            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_2007_pd.html

19    US EPA (20 lOa).  Quantitative Risk Assessment for Paniculate Matter - Second External Review Draft. Office of
20            Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC.
21            EPA-452/P-10-001.  February 2010. Available:
22            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_index.html.

23    Yang C-Y; Chen Y-S; Yang C-H; Ho S-C (2004). Relationship between ambient air pollution and hospital
24            admissions for cardiovascular diseases in Kaohsiung, Taiwan. J Toxicol Environ Health A Curr Iss, 67:
25            483-493.

26    Yang CY; Cheng MH; Chen CC (2009). Effects of Asian Dust Storm Events on Hospital Admissions for Congestive
27            Heart Failure in Taipei, Taiwan. J Toxicol Environ Health A Curr Iss, 72: 324-328.

28    Zanobetti A; Schwartz J. (2009). The effect of fine and coarse paniculate air pollution on mortality: A national
29            analysis.  Environ Health Perspect, 117:  1-40.(2008).
              Draft Do Not Quote or Cite            3-42                           March 2010

-------
 1     4      REVIEW OF THE SECONDARY STANDARDS FOR VISIBILITY-
 2                                     RELATED EFFECTS

 3           This chapter presents preliminary staff conclusions with regard to the adequacy of the
 4    current suite of secondary PM standards to protect against PM-related visibility impairment as
 5    well as the alternative secondary PM standards that are appropriate for consideration in this
 6    review.  Our assessment of these issues is framed by a series of key policy-relevant questions,
 7    which expand upon those presented in the Integrated Review Plan (IRP) (US EPA, 2008a)  at the
 8    outset of this review.  The answers to these questions will inform decisions on whether, and if so
 9    how, to revise the current suite of secondary PM standards for the purpose of providing
10    appropriate protection from PM-related visibility impairment.
11           In presenting preliminary staff conclusions on alternative secondary standard that are
12    appropriate for consideration, we note that the final decision is largely a public welfare policy
13    judgment.  A final decision must draw upon scientific information and analyses about PM-
14    related visibility impairment and related impacts on public welfare, as well as judgments about
15    how to deal with the range of uncertainties that are inherent  in the scientific evidence and
16    analyses. Our approach to informing these judgments is discussed more fully below.  This
17    approach is consistent with the requirements of the NAAQS provisions of the Act and with how
18    EPA and the courts have historically interpreted the Act. These provisions require the
19    Administrator to establish secondary standards that, in the Administrator's judgment, are
20    requisite to protect public welfare from any known or anticipated adverse effects associated with
21    the presence of the pollutant in the ambient air.  In so doing, the Administrator seeks to establish
22    standards that are neither more nor less stringent than necessary for this purpose. The Act does
23    not require that secondary standards be set at a zero-risk level, but rather at a level that avoids
24    unacceptable public welfare impacts.
25           Information on the approaches used to set the  secondary PM standards in past reviews as
26    well as our current approach for this review are presented in section 4.1. Our preliminary
27    conclusions regarding the adequacy of the current suite of secondary PM standards to protect
28    against PM-related visibility impairment are presented in section 4.2.  Section 4.3 presents  our
29    preliminary conclusions with respect to alternative PM standards by focusing on each of the
30    basic elements of the standards:  pollutant indicator (section 4.3.1), averaging time (section
31    4.3.2), and level and form (section 4.3.3).  The performance of alternative standards, with a focus
32    on the uniformity of protection from visibility impairment afforded by the alternative standards,
33    is evaluated in  section 4.3.4.  Section 4.4 summarizes all preliminary staff conclusions on the
34    secondary PM  standards for visibility protection. The next draft of this chapter will conclude
35    with an initial overview of areas of key uncertainties and suggested future research areas and
36    data collection efforts (section 4.5)

                                                4-1
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    4.1   APPROACH
 2           Staffs approach for reviewing the current secondary PM standards builds upon and
 3    broadens the approaches used in previous PM NAAQS reviews. We first present a brief
 4    summary of the approaches used to establish and review secondary PM standards in the last two
 5    reviews of the PM NAAQS (section 4.1.1).  Recent litigation on the 2006 standards has resulted
 6    in the remand of the secondary annual and 24 hour PM25 NAAQS to EPA as discussed in section
 7    4.1.2. Our current approach for evaluating the secondary PM2.5 standards using both evidence-
 8    and impact assessment-based considerations is outlined in section 4.1.3.

 9    4.1.1   Approaches Used in Previous Reviews
10           The original suite of PM2 5 standards was established in 1997 and revisions to those
11    standards were made in 2006. The approaches used in making final decisions on secondary
12    standards in those reviews, as well as the current review, utilize different ways to consider the
13    underlying body of scientific evidence.  They also reflect an evolution in our understanding of
14    the multi-faceted nature of the public welfare visibility effect, from a more narrow focus on
15    Class I area visibility to non-Class I  area visibility, including urban areas, and public perception,
16    valuation and impacts on personal comfort and well being.
17           4.1.1.1  Review Completed  in 1997
18           In 1997, EPA revised the identical primary and secondary PMi0 NAAQS in part by
19    establishing new identical primary and secondary PM2.5 standards.  In revising the secondary
20    standards, EPA recognized that PM  produces adverse effects on visibility and that impairment of
21    visibility was experienced throughout the U.S., in multi-state regions, urban areas, and remote
22    mandatory Class I Federal areas alike. However, in considering an appropriate level for a
23    secondary standard to address adverse effects of PM on visibility,  EPA concluded that the
24    determination of a single national level was complicated by regional differences in several
25    factors that  influence visibility such  as background and current levels of PM, composition  of PM,
26    and average relative humidity. Variations in these factors could thus result in situations where
27    attaining even a low concentration of fine particles might or might not provide adequate
28    protection, depending on these factors.  The EPA also determined that there was insufficient
29    information at that time to establish  a level for a national secondary standard that would
30    represent a threshold above which visibility conditions would always be adverse and below
31    which visibility conditions would always be acceptable.
32           Based on these considerations, EPA assessed potential visibility improvements in urban
33    areas and on a regional scale that would result from attainment of the new primary standards for
34    PM2 5.  The agency concluded that the spatially averaged form of the annual PM2 5 standard was
35    well suited to the protection of visibility, which involves effects of PM throughout an extended

                                                4-2
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1    viewing distance across an urban area. Based on air quality data available at that time, many
 2    urban areas in the northeast, midwest, and southeast, as well as Los Angeles, were expected to
 3    see perceptible improvement in visibility if the annual PM2 5 primary standard was attained. The
 4    EPA also concluded that in some areas attainment of the 24-hour PM2.5 standard would be
 5    expected to reduce, to some degree, the number and intensity of "bad visibility" days, i.e., the
 6    20% of days having the greatest impairment over the course of a year.
 7           Having concluded that attainment of the annual and 24-hour PM2.5 primary standards
 8    would lead to visibility improvements in many eastern and some western urban areas, EPA also
 9    considered whether these standards could provide potential improvements to visibility on a
10    regional scale.  Based on information available at the time, EPA concluded that attainment of
11    PM2.s secondary standards set identical to the primary standards would be expected to result in
12    visibility improvements in the eastern U.S. at both urban and regional scales, but little or no
13    change in the western U.S., except in and near certain urban areas.
14           The EPA then considered the potential effectiveness of a regional haze program, required
15    by sections 169 A and 169B of the Act1 to address the widespread, regionally uniform type  of
16    haze caused by a multitude of sources, to address those effects of PM on visibility that would not
17    be addressed through attainment of the primary PM2.5  standards.  The structure and requirements
18    of sections 169A and 169B of the Act provide for visibility protection programs that can be more
19    responsive to the factors contributing to regional differences in visibility than can programs
20    addressing a nationally applicable secondary NAAQS. The regional haze visibility goal was
21    more protective than  a secondary NAAQS since the goal addresses any man-made impairment
22    rather than just impairment at levels determined to be adverse.  Thus, an important factor
23    considered in the 1997 review was whether a regional haze program, in conjunction with
24    secondary standards set identical to the suite of PM2.s primary standards, would provide
25    appropriate protection for visibility in non-Class I areas. The EPA concluded that the two
26    programs and associated control strategies should provide  such protection due to the regional
27    approaches needed to manage emissions of pollutants that  impair visibility in many of these
28    areas.
29       For these reasons, EPA concluded that a national regional haze program, combined with a
30    nationally applicable level of protection achieved through  secondary PM2 5 standards set
31    identical to the primary PM2.5 standards, would be more effective in addressing regional
32    variations in the adverse effects  of PM2 5 on visibility than establishing national secondary
      1 In 1977, Congress established as a national goal' 'the prevention of any future, and the remedying of any existing,
      impairment of visibility in mandatory Class I Federal areas which impairment results from manmade air pollution",
      section 169A(a)(l) of the Act. The EPA is required by section 169A(a)(4) of the Act to promulgate regulations to
      ensure that "reasonable progress" is achieved toward meeting the national goal.
                                                 4-3
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    standards for PM at lower levels than the primary PM2.5 standards. The EPA further recognized
 2    that people living in certain urban areas may place a high value on unique scenic resources in or
 3    near these areas, and as a result might experience visibility problems attributable to sources that
 4    would not necessarily be addressed by the combined effects of a regional haze program and
 5    PM2.5 secondary standards. The EPA concluded that in such cases, State or local regulatory
 6    approaches, such as past action in Colorado to establish a local visibility standard for the City of
 7    Denver, would be more appropriate and effective in addressing these special situations because
 8    of the localized and unique characteristics of the problems involved. Visibility in an urban area
 9    located near a mandatory Class I Federal  area could also be improved through State
10    implementation of the current visibility regulations, by which emission limitations can be
11    imposed on a source or group of sources found to be contributing to "reasonably attributable''
12    impairment in the mandatory Class I Federal area.
13          Based on these considerations, EPA set secondary PM2.5 standards identical to the
14    primary PM2.5 standards, in conjunction with a regional haze program under sections 169A and
15    169B of the Act, as the  most appropriate and effective means of addressing the welfare effects
16    associated with visibility impairment. Together, the two programs and associated control
17    strategies were expected to provide appropriate protection against the effects of PM on visibility
18    and enable all regions of the country  to make reasonable progress toward the national visibility
19    goal.
20          4.1.1.2 Review Completed in 2006
21          In 2006, EPA revised the secondary PM standards by once again making them identical
22    to the revised primary standards. The EPA's decision to revise the secondary PM standards
23    reflected a number of new developments  and sources of information that had occurred and/or
24    become available following the 1997 review. First, EPA promulgated a regional haze program
25    in 1999 (65 FR 35713) which required States to establish goals for improving visibility in Class I
26    areas and to adopt control strategies to achieve these goals.  Second, extensive new information
27    from visibility and fine  particle monitoring networks had become available, allowing for updated
28    characterizations of visibility trends and PM levels in urban areas, as well as Class I areas.
29    These new data allowed EPA to better characterize visibility impairment in urban areas and the
30    relationship between visibility and PM2 5 concentrations. Finally, additional studies in the U.S.
31    and abroad provided the basis for the establishment of standards and programs to address
32    specific visibility concerns in a number of local areas. These studies (e.g., in Denver, Phoenix,
33    British Columbia) produced reasonably consistent results in terms of the visual ranges found to
34    be generally acceptable by the participants in the various studies. These studies utilized
35    photographic representations of visibility impairment, which were useful in characterizing the
36    nature of particle-induced haze.  Based largely on this information, the Administrator concluded

                                                 4-4
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1    that it was appropriate to revise the secondary PM standards to provide increased protection from
 2    visibility impairment principally in urban areas, in conjunction with the regional haze program
 3    for protection of visual air quality in Class I areas.
 4           In so doing, the Administrator recognized that PM-related visibility impairment is
 5    principally related to fine particle levels, such that it was appropriate to focus the review on
 6    whether the current secondary PM2.5 standards should be revised. The Administrator also
 7    recognized that perception of visibility impairment is most directly related to instantaneous
 8    levels of visual air quality, such that in considering whether the current suite of secondary
 9    standards would provide the appropriate degree of protection, he concluded that it was
10    appropriate to revise just the 24-hour secondary PM2.5 standard to provide requisite protection.
11           The EPA then considered whether PM2 5 remained the appropriate indicator for a
12    secondary standard to protect visibility primarily in urban areas. One of the key issues in the
13    1997 review was the extent to which the differences in humidity between East and West
14    complicated  the establishment of a nationally uniform PM2.5 secondary standard for urban areas.
15    With the substantial addition to the air quality and visibility data made possible by the national
16    urban PM2.5 monitoring networks, an analysis conducted for the 2006 review found that, in urban
17    areas, visibility levels showed far less difference between eastern and western regions on a 24-
18    hour or shorter time basis than implied by the largely non-urban data available in the 1997
19    review. In analyzing how well PM2 5 concentrations correlated with visibility in urban locations
20    across the U.S., the 2005 Staff Paper concluded that clear correlations existed between 24-hour
21    average PM2.5 concentrations and reconstructed light extinction, which is directly related to
22    visual range. These correlations were similar in the eastern and western regions of the U.S.
23    Particles in the coarse mode generally contributed only marginally to visibility impairment in
24    urban areas.  Further, because hygroscopic components of fine particles, in particular sulfates
25    and nitrates,  contribute disproportionately to visibility impairment under high humidity
26    conditions, these correlations were less influenced by relative humidity and more consistent
27    across regions  when PM2.5 concentrations are averaged over shorter, daylight time periods (e.g.,
28    4 to 8 hours) when relative humidity was generally lower and less variable.  The 2005 Staff
29    Paper noted that a standard set at any specific PM2.5 concentration would necessarily result in
30    visual ranges that vary somewhat in urban areas across the country, reflecting the variability in
31    the correlations between PM2.5 concentrations and light extinction.  Thus, the 2005 Staff Paper
32    concluded that it was appropriate to use PM2.5 as an indicator for standards to address visibility
33    impairment in urban areas, especially when the indicator is defined for a relatively short period
34    (e.g., 4 to 8 hours) of daylight hours. Based on their review of the Staff Paper, most CASAC
35    Panel members also endorsed such a PM2.5 indicator for a secondary standard to address
36    visibility impairment (Henderson, 2005a). Based on the above considerations, the Administrator

                                                 4-5
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1    concluded that PM2.5 should be retained as the indicator for fine particles as part of a secondary
 2    standard to address visibility protection, in conjunction with averaging times from 4 to 24 hours.
 3           In considering what level of protection against PM-related visibility impairment would be
 4    appropriate, the Administrator took into account the results of the public perception and attitude
 5    surveys in the U.S. and Canada, State and local visibility standards within the U.S., and visual
 6    inspection of photographic representations of several urban areas across the U.S. In the
 7    Administrator's judgment, these sources provided useful but still quite limited information on the
 8    range of levels appropriate for consideration in setting a national visibility standard primarily for
 9    urban areas, given the generally subjective nature of the public welfare  effect involved. Based
10    on photographic representations of varying levels of visual air quality, public perception studies,
11    and local and State visibility standards, the 2005 Staff Paper concluded that 30 to 20 ug/m3 PM2.5
12    represented a reasonable range for a national visibility standard primarily for urban areas, based
13    on a sub-daily averaging time.  The upper end of this range was below the levels at which
14    illustrative scenic views are significantly obscured, and the lower end was around the level at
15    which visual air quality generally appeared to be good based on observation of the illustrative
16    views.  This concentration range generally corresponded to median visual ranges in urban areas
17    within regions  across the U.S. of approximately 25 to 35 km, a range which was bounded above
18    by the visual range targets selected in specific areas where State or local agencies placed
19    particular emphasis on protecting visual air quality. In considering a reasonable range  of forms
20    for a PM2.5 standard within this range of levels, the 2005 Staff Paper concluded that a
21    concentration-based percentile form was appropriate, and that the upper end of the range of
22    concentration percentiles should be consistent with the 98th percentile used for the primary
23    standard and that the lower end of the range should be the 92nd percentile, which represented the
24    mean of the distribution of the 20 percent most impaired days, as targeted in the regional haze
25    program. While recognizing that it was difficult to select any specific level and form based on
26    then currently available information (Henderson, 2005a), the CASAC Panel was generally in
27    agreement with the ranges of levels and forms  presented in the 2005 Staff Paper.
28           The Administrator also considered the level of protection that would be afforded by the
29    proposed suite of primary PM2 5 standards (71  FR 2681), on the basis that although significantly
30    more information was available than in the 1997 review concerning the relationship between fine
31    PM levels  and visibility across the country, there was still little available information for use in
32    making the relatively subjective value judgment needed in selecting the appropriate degree of
33    protection to be afforded by such a standard.  In so doing, the Administrator compared the extent
34    to which the proposed suite of primary standards would require areas across the country to
35    improve visual air quality with the extent of increased protection likely  to be afforded by a
36    standard based on a sub-daily averaging time.  Based on such an analysis, the Administrator

                                                 4-6
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    observed that the predicted percent of counties with monitors not likely to meet the proposed
 2    suite of primary PM2.5 standards was actually somewhat greater than the predicted percent of
 3    counties with monitors not likely to meet a sub-daily secondary standard with an averaging time
 4    of 4 daylight hours, a level toward the upper end of the range recommended in the 2005 Staff
 5    Paper, and a form within the recommended range. Based on this comparison, the Administrator
 6    concluded that revising the secondary 24-hour PM2.5 standard to be identical to the proposed
 7    revised primary PM2 5 standard (and retaining the current annual secondary PM2 5 standard) was a
 8    reasonable policy approach to addressing visibility protection primarily in urban areas. In
 9    proposing this approach, the Administrator also solicited comment on a sub-daily (4- to 8-hour
10    averaging time) secondary PM2 5 standard (71 FR 2675-2781).
11           In commenting on EPA's proposal, the CAS AC requested that a sub-daily standard to
12    protect visibility be favorably reconsidered (Henderson, 2006).  The CASAC noted three
13    cautions regarding the Agency's proposed reliance on a secondary PM2 5 standard identical to the
14    proposed 24-hour primary PM2.5 standard:  (1) PM2.5 mass measurement is a better indicator of
15    visibility impairment  during daylight hours, when relative humidity is generally low; the sub-
16    daily standard more clearly matches the nature of visibility impairment, whose adverse effects
17    are most evident during the daylight hours; using a 24- hour standard as a proxy introduces error
18    and uncertainty in protecting visibility; and sub-daily standards are used for other NAAQS and
19    should be the focus for visibility; (2) CASAC and its monitoring subcommittees have repeatedly
20    commended EPA's initiatives promoting the introduction of continuous and near-continuous PM
21    monitoring, and expanded deployment of continuous PM2 5 monitors is consistent with setting a
22    sub-daily standard to  protect visibility; (3) The analysis showing a similarity between
23    percentages of counties not likely to meet what they considered to be a lenient 4- to 8-hour
24    secondary standard and a secondary standard identical to the proposed 24-hour primary standard
25    was a numerical coincidence that was not indicative of any fundamental relationship between
26    visibility and health.  The CASAC Panel further stated that "visual air quality is substantially
27    impaired at PM2 5 concentrations of 35 ug/m3" and that "it is  not reasonable to have the visibility
28    standard tied to the health standard, which may change in ways that make it even less appropriate
29    for visibility concerns."
30           In reaching a final decision, the Administrator focused on the limitations in the available
31    hourly air quality data and in the studies of public perception  and attitudes regarding the
32    acceptability of various degrees of visibility impairment in urban areas, as well as on the
33    subjective nature of the public welfare judgment required. In so doing, the Administrator
34    concluded that caution was warranted in establishing a distinct secondary standard for visibility
35    impairment and that the available information did not warrant adopting a secondary standard that
                                                4-7
     Draft - Do Not Quote or Cite                                                  March 2010

-------
 1    would provide either more or less protection against visibility impairment in urban areas than
 2    would be provided by secondary standards set equal to the proposed primary PM2.5 standards.

 3    4.1.2  Remand of Secondary PM2.s Standards
 4          Several parties filed petitions for review following promulgation of the revised PM
 5    NAAQS in 2006. These petitions challenged several aspects of the final rule including EPA's
 6    decision to set the secondary NAAQS for fine PM, which protect the public welfare from
 7    adverse visibility effects, at the same level as the primary NAAQS, which protect public health
 8    42U.S.C. 7607(d)(9).
 9          On judicial review, the D.C. Circuit remanded for reconsideration the secondary NAAQS
10    for fine PM to EPA because the Agency's decision was unreasonable and contrary to the
11    requirements of 42 U.S. C. section 7409 (b)(2). The petitioners argued that EPA's decision
12    lacked a reasoned basis. First, they asserted, EPA never determined what level of visibility was
13    "requisite to protect the public welfare" 42 U.S.C. section 7409(b)(2). They argued that EPA
14    unreasonably rejected the target level of protection provided by its staff, while failing to provide
15    a target level of its own. The D.C. Circuit court concurred stating "the EPA's failure to identify
16    such a level when deciding where to set the level of air quality required by the revised secondary
17    fine PM NAAQS is contrary to the statute and therefore unlawful. Furthermore, the failure to set
18    any target level of visibility protection deprived the EPA's decisionmaking of a reasoned basis."
19    American Farm Bureau Federation v. EPA, 559 F. 3d 512, (D.C. Cir. 2009).
20          Second, the petitioners challenged EPA's method of comparing the protection expected
21    from potential standards. They contend that the EPA relied on  a meaningless numerical
22    comparison, ignored the effect of humidity on the usefulness of a standard using a daily
23    averaging time, and unreasonably concluded that the primary standards would achieve a level of
24    visibility roughly equivalent to the level the EPA staff and CASAC deemed "requisite to protect
25    the public welfare."  Again, the court found that EPA's equivalency analysis failed  on its own
26    terms. The same table showing the alternative secondary standard used for comparison to the
27    alternative primary to show equivalency also included six other standards within the
28    recommended CASAC range that would be more "protective" under EPA's definition than the
29    primary standards. Two thirds of the potential standards within the CASAC's recommended
30    range would be substantially more protective than the primary standards. The EPA failed to
31    explain why it looked only at one of the few potential standards that would be less protective-and
32    only slightly so- than the primary standards. More fundamentally, however, EPA's equivalency
33    analysis demonstrated nothing about the relative protection offered by the different standards.
34    The tables simply offer no valid information about the relative visibility protection provided by
35    the standards.

                                                4-8
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1          Finally, the Staff Paper made clear that a visibility standard using a daily averaging time
 2    will be confounded by regional differences in humidity.  The EPA acknowledged this problem,
 3    yet did not address this issue at all in concluding that the primary standards would be sufficiently
 4    protective of visibility.  Therefore, the court granted the petition for review and remanded for
 5    reconsideration the secondary  PM NAAQS. American Farm Bureau Federation, 559 F. 3d at
 6    520-24.

 7    4.1.3  Current Approach
 8          The staffs approach in this review broadens the general approaches used in the last two
 9    PM NAAQS reviews by utilizing, to the extent available, enhanced tools, methods, and data to
10    more comprehensively characterize visibility impacts.  As such,  staff is taking into account both
11    evidence-based and impact-based considerations to inform staff conclusions related to the
12    adequacy of the current secondary standards and alternative standards that are appropriate for
13    consideration in this review. In so doing, we are seeking to provide as broad an array of options
14    as is supportable by the available information, recognizing that the selection of a specific
15    approach to reaching final decisions on the secondary PM2 5 standards will reflect the judgments
16    of the Administrator as to what weight to place on the various approaches and types of
17    information presented in the final PA.
18          For the purposes of this first draft Policy Assessment  (PA), staff has drawn from the
19    qualitative evaluation of all studies evaluated  in the Integrated Science Assessment for
20    Paniculate Matter (FinalReport) (ISA, USEPA, 2009a). The discussions presented in this
21    chapter consider the extensive new air quality and source apportionment information available
22    from the regional planning organizations, long-standing evidence of PM effects on visibility, and
23    public preference studies from four urban areas, as discussed in chapter  9 of the ISA,  as well as
24    the integration of evidence across disciplines presented in chapter 2 of the ISA. In addition,
25    limited information that has become available regarding the characterization of public
26    preferences in urban areas has provided some new perspectives on the usefulness of this
27    information in informing the selection of target levels of urban visibility protection.  On these
28    bases, we are again focusing our assessments  in this review on visibility conditions in urban
29    areas.
30          Our preliminary conclusions reflect our understanding of both evidence-based and
31    impact-based considerations to inform two overarching questions related to: (1) the adequacy of
32    the current suite of PM2.5 standards and (2) what potential alternative standards, if any, should be
33    considered in this review to provide appropriate protection from PM-related visibility
34    impairment.  In addressing these broad questions, we have organized the discussions below
35    around a series of more specific questions reflecting different aspects of each overarching

                                                4-9
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    question. When evaluating the visibility protection afforded by the current or any alternative
 2    standards considered, we have taken into account the four basic elements of the NAAQS (e.g.,
 3    indicator, averaging time, level, and form).
 4          We believe that the approach outlined above, when presented in the final PA, will
 5    provide a comprehensive basis to help inform the judgments required of the Administrator in
 6    reaching decisions about the current and potential alternative secondary PM standards for the
 7    purpose of providing appropriate protection from PM-related visibility impairment and in
 8    responding to the remand of the 2006 secondary PM2.5 standards.

 9    4.2   ADEQUACY OF CURRENT STANDARDS
10          In considering the adequacy of the current suite of PM2.5 standards, staff addresses the
11    following overarching question:
12
13
14
      Does the currently available scientific evidence and visibility impact information, as
reflected in the ISA and UFVA, support or call into question the adequacy of the visibility
           protection afforded by the current suite of fine particle standards?
15          To inform the answer to this overarching question, we have posed a series of more
16    specific questions to aid in considering the currently available scientific evidence and the results
17    of recent quantitative visibility impact analyses in a policy-relevant context, as discussed below.
18    In considering the scientific and technical information, we reflect upon both the information
19    available in the last review and information that is newly available since the last review as
20    assessed and presented in the ISA and UFVA (US EPA, 2009a; US EPA,  201 Ob).

21    4.2.1  Evidence-based Considerations
22          In reviewing the adequacy of the current suite of PM2.5 standards, we have taken into
23    account evidence-based considerations, primarily as presented in the ISA, by considering causal
24    inference, impacts on susceptible populations, and whether visibility effects have been  observed
25    in urban areas that would likely meet the current standards.
26    •   To what extent does the newly available scientific evidence strengthen or call into
27       question evidence of associations between ambient fine particle exposures and  visibility
28       effects?
29          New research conducted by regional planning organizations in support of the Regional
30    Haze Rule, as discussed in chapter 9 of the ISA, continues to support and  refine our
31    understanding of the nature of the PM visibility effect and the source contributions to that effect
32    in rural and remote locations. Additional byproducts of this research include new insights
33    regarding the regional source contributions to urban visibility and better characterization of the
34    urban excess that occurs in many cities above regional background.  Ongoing urban PM2.5

                                               4-10
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1    speciated and mass monitoring has produced new information that has allowed for updated
 2    characterization of visibility trends and current levels in urban areas.  Information from both of
 3    these sources of PM data, while useful, has not however changed the fundamental and long
 4    understood science characterizing the contribution of PM, especially fine particles, to visibility
 5    impairment from the last review.  This science, briefly summarized below, provides the basis for
 6    the ISA designation of the relationship between PM and visibility impairment as causal.
 7           Visibility impairment is caused by the scattering and absorption of light by suspended
 8    particles and gases in the atmosphere. The combined effect of light scattering and absorption by
 9    both particles and  gases is characterized as light extinction, (i.e. the fraction of light that is
10    scattered or absorbed per unit of distance in the atmosphere). Light extinction is measured in
11    units of I/distance, which is often expressed in the technical literature as l/(million meters) or
12    inverse megameters (abbreviated Mm"1). When PM is present in the air, its contribution to light
13    extinction typically greatly exceeds that of gases.
14           The amount of light extinction contributed by PM depends on the particle size
15    distribution and composition, as well as its concentration. If details of the ambient particle size
16    distribution and composition (including the  mixing of components) are known, Mie theory can
17    be used to accurately calculate PM light extinction (ISA chapter 9).  However, routine
18    monitoring rarely includes measurements of particle  size and composition information with
19    sufficient detail for such calculations. A much simpler algorithm  can be used to estimate light
20    extinction using routinely monitored fine particle (PM2.5) speciation and coarse particle mass
21    (PMio-2.s) data, plus relative humidity information needed to estimate the contribution by liquid
22    water in solution with hygroscopic PM components (ISA section 9.2.2.2 and UFVA chapter 3).
23           The concentration of each of the major aerosol components is multiplied by a dry
24    extinction efficiency value  and for the hygroscopic components (e.g., ammoniated sulfate and
25    ammonium nitrate) an additional multiplicative term to account for the water growth to estimate
26    that components contribution to light extinction. Both the dry extinction efficiency and water
27    growth terms are developed by some combination of empirical assessment and theoretical
28    calculation using typical particle size distributions associated with each of the major aerosol
29    components, and they are evaluated by comparing the algorithm estimates of light extinction
30    with coincident optical measurements. Summing the contribution  of each component gives the
31    estimate of total light extinction. The most commonly used of these is referred to as the
32    IMPROVE algorithm because it was developed specifically to use the IMPROVE aerosol
33    monitoring data and was evaluated using IMPROVE optical measurements at the subset of sites
34    that make those measurements (Malm et al., 1994). The formula for the traditional IMPROVE
35    algorithm is shown below.
36

                                                4-11
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1         bext -   3 x f(RH) x [Sulfate]
 2              + 3 x f(RH) x [Nitrate]
 3              + 4 x [Organic Mass]
 4              + 10 x  [Elemental Carbon]
 5              + 1 x [Fine Soil]
 6              + 0.6 x [Coarse Mass]
 7              +10
 8
 9          Light extinction (£ext) is in units of Mm"1, the mass concentrations of the components
10    indicated in brackets are in ug/m3, andf(RH) is the unitless water growth term that depends on
11    relative humidity.  The dry extinction efficiency for particulate organic mass is larger than those
12    for particulate SO42  and nitrate principally because the density of the dry inorganic compounds
13    is higher than that assumed for the PM organic mass components.  Since IMPROVE does not
14    include ammonium ion monitoring, the assumption is made that all SO42 is fully neutralized
15    ammonium sulfate and all nitrate is assumed to be ammonium nitrate. Though often reasonable,
16    neither assumption is always true (see Section 9.2.3.1). In the eastern U.S.  during the summer
17    there is insufficient ammonia in the atmosphere to neutralize the SO42 fully.  Fine particle
18    nitrates can include sodium or calcium nitrate, which are the fine particle fraction of generally
19    much coarser particles due to nitric acid interactions with sea salt at near-coastal areas (sodium
20    nitrate) or nitric acid interactions with calcium carbonate in crustal aerosol (calcium nitrate).
21    Despite the simplicity of the algorithm, it performs reasonably well and  permits the contributions
22    to light extinction  from each of the major components  (including the water  associated with the
23    SO42 and nitrate compounds) to be separately approximated.
24          Thef(RH) terms inflate the particulate SO42  and nitrate light scattering for high relative
25    humidity conditions.  For relative humidity below 40% thef(RH) value is 1, but it increases to 2
26    at -66%, 3 at -83%, 4 at -90%,  5 at -93% and 6 at -95% relative humidity.  The result is that
27    both particulate SO42 and nitrate are more efficient per unit mass than any  other aerosol
28    component for relative humidity above -85% where its total light extinction efficiency exceeds
29    the 10m2/g associated with EC. Based on this algorithm, particulate  SO42  and nitrate are
30    estimated to have comparable light extinction efficiencies (i.e., the same dry extinction
31    efficiency andf(RH) water growth terms), so on a per unit mass concentration basis at any
32    specific relative humidity they are treated as equally effective contributors to visibility effects.
33          Inspection  of the PM component-specific terms in the simple algorithm shows that most
34    of the PM2 5 components contribute 5 times or more light extinction than a similar concentration
35    of PMio-2.5- We also know that particles with hygroscopic components (e.g.  particulate sulfate
36    and nitrate) contribute more light extinction at higher relative humidity than at lower relative
37    humidity because they change size in the atmosphere in response to ambient relative humidity
38    conditions.  PM containing elemental or black carbon absorbs light as well  as scattering it,
                                                4-12
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    making it the component with the greatest light extinction contributions per unit of mass
 2    concentration, except for the hygroscopic components under high relative humidity conditions.
 3           As a result of better characterization of PM in rural and remote locations, refinements in
 4    the original IMPROVE algorithm have been made to better account for the aging of the organic
 5    aerosols that occurs during transport to these more distant areas, and to add a component for sea
 6    salt.  Since urban areas contain a higher proportion of newly emitted aerosols that have not aged,
 7    the assessments done in support of this review have relied on the original IMPROVE algorithm.
 8           As mentioned above, particles are not the only contributor to ambient visibility
 9    conditions. Light scattering by gases also occurs in ambient air.  Under pristine atmospheric
10    conditions, naturally occurring gases such as N2 and O2, cause what is known as Rayleigh
11    scattering. Rayleigh scattering, which depends on the density of air as a function primarily of the
12    elevation above sea level, can be treated as a site-dependent constant.  Rayleigh contribution to
13    light extinction is only significant under pristine conditions. The only other commonly occurring
14    atmospheric gas to appreciably absorb light in the visible spectrum is NO2.  NO2 forms in the
15    atmosphere from NO emissions associated with combustion processes. These combustion
16    processes also emit PM at  levels that generally contribute much higher light extinction than the
17    NO2 (i.e. NO2 absorption is generally less than ~ 5% of the light extinction, except where
18    emission controls remove most of the PM prior to releasing the remaining gases to the
19    atmosphere). The remainder of this document focuses on the contribution of PM, which is
20    typically much greater than that of gases, to ambient light extinction, unless otherwise specified.
21    •  To what extent does the available evidence inform our understanding of the temporal
22       nature of the PM visibility effect, including relevant exposure periods, associated
23       atmospheric conditions, and diurnal  patterns of exposure?
24    Diurnal Periods of Interest
25           Typically, we think of visibility associated with daytime periods. We recognize,
26    however, that PM light extinction behaves the same at night as during the day, enhancing the scattering
27    of anthropogenic light, contributing to the "skyglow" within and over populated areas, adding to the total
28    sky brightness, and contributing to the reduction in contrast of stars against the background. These effects
29    produce the visual result of a reduction of the number of visible stars and the disappearance of diffuse or
30    subtle phenomena such as the Milky Way. The extinction of starlight is a secondary and minor effect also
31    caused by increased PM scattering and absorption.
32           However, there are significant and important differences between daytime and nighttime visual
33    environments that potentially make the nighttime period inappropriate to address at this time. First,
34    daytime visibility has dominated the attention of those who have studied the visibility effects of air
35    pollution. As a result, little research has been conducted on nighttime visibility and the state of the
36    science is not yet comparable to that associated with daytime visibility impairment. Second, in addition

                                                 4-13
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    to air pollution, nighttime visibility is affected by the addition of light into the sight path from numerous
 2    sources, including anthropogenic light sources such as artificial outdoor lighting, which varies
 3    dramatically across space, and natural sources including the Moon, planets, and stars.  Light sources and
 4    ambient conditions are typically five to seven orders of magnitude dimmer at night than in sunlight.
 5    Moonlight, like sunlight, introduces light throughout an observer's sight path at a constant angle. On the
 6    other hand, dim starlight emanates from all over the celestial hemisphere while artificial lights are
 7    concentrated in cities and illuminate the atmosphere from below. These different light sources will yield
 8    variable changes in visibility as compared to what has been established for the daytime scenario, in which
 9    a single source, the sun, is by far the brightest source of light. Third, the human psychophysical response
10    (e.g., how the human eye sees and processes visual stimuli) at night is expected to differ (ISA, section
11    9.2.2).
12           Given the above, we do not believe that the science is available at this time to support
13    adequate characterization of nighttime PM light extinction effects. In addition, we are not aware
14    of preference or valuation studies providing information on public preferences for nighttime
15    visual air quality (VAQ). Thus, we limit our consideration of PM visibility impacts to all
16    daylight hours only.
17    Exposure Durations of Interest
18           Very little is known about the role exposure duration plays in determining the
19    acceptability or unacceptability of a given level of VAQ on the public welfare.  We do know
20    from preference and/or valuation studies that atmospheric visibility conditions can be quickly
21    assessed and preferences determined. These studies show that a momentary glance at an image
22    of a scene (i.e. less than a minute) is enough for study participants to consistently judge the
23    acceptability or unacceptability of the viewed visual air quality conditions. Outside these
24    controlled settings, we are unaware  of any studies that characterize the extent to which different
25    frequencies and durations of exposure to visibility conditions contribute to the degree of public
26    welfare impact that occurs.
27           In some circumstances, such as infrequent visits to scenic vistas in natural or urban
28    environments, people are motivated specifically to take the opportunity to view a valued scene
29    and are likely to do so for many minutes to hours to  appreciate various aspects of the vista they
30    choose to view. In such circumstances, the viewer may consciously evaluate how the VAQ  at
31    that time either  enhances or diminishes the experience/view.  However, the public also has many
32    more opportunities to notice visibility conditions on a daily basis in settings/surroundings
33    associated with performing daily routines (e.g. during commutes, while walking the dog, or
34    when taking out the recyclables).  These scenes, whether iconic or generic, may not be
35    consciously viewed for their scenic value, but their VAQ can still affect a person's sense of
36    wellbeing. Research has demonstrated that people are emotionally affected by  low VAQ, that

                                                 4-14
      Draft - Do Not Quote or Cite                                                    March 2010

-------
 1    perception of pollution is correlated with stress, annoyance and symptoms of depression, and that
 2    VAQ is deeply intertwined with a "sense of place," affecting peoples sense of the desirability of
 3    a neighborhood (ISA section 9.2.4). Though we do not know the extent to which these
 4    emotional effects are linked to different periods of exposure to poor VAQ, providing additional
 5    protection against even short term exposures to levels of VAQ considered unacceptable would be
 6    expected to provide some degree of improvement in the public's "sense of wellbeing".
 7           Some people have mostly intermittent opportunities on a daily basis (e.g. during morning
 8    and/or afternoon commutes) to experience ambient visibility conditions as they spend much of
 9    their time indoors without access to windows. For such people a view of poor VAQ during their
10    morning commute may provide their perception of the day's visibility conditions until the next
11    time they venture outside during daylight hours later or perhaps the next day. Other people have
12    exposure to visibility conditions throughout the day, so that a day with multiple hours of
13    visibility impairment would likely be judged a having a greater impact on their wellbeing than a
14    day with just one such hour.
15           We have no information or studies  on the fraction of the public that has only one or a
16    few opportunities to experience visibility during the day, or information or studies on the
17    duration of the effect on wellbeing from exposure to different durations of poor VAQ conditions.
18    However, it is logical to conclude that people with limited opportunities to experience visibility
19    conditions on a daily basis would receive the entire impact of the day's VAQ based on the
20    visibility conditions that occur during the short time period when they can see it.  On the basis of
21    this rationale, the segment of the population with infrequent access to visibility could be
22    characterized as a sensitive subpopulation. Since this group could be affected on the basis of
23    observing VAQ conditions for periods as short as one hour or less, we believe it is appropriate to
24    consider characterizing PM visibility conditions in terms of the worst or maximum hourly value
25    during daylight hours for each day for purposes of evaluating the adequacy of the current suite of
26    secondary standards.
27           For another group of observers, those who have access to visibility conditions often or
28    continuously throughout the day, the entire impact of the day's visibility conditions may be
29    based on the varying visibility conditions they observe throughout the day. For this group, it
30    might be more appropriate to evaluate the adequacy of the current suite of secondary standards in
31    terms of all daylight hours in the day.
32           While both the maximum daily and all hour averaging times were assessed in the UFVA
33    for the PM light extinction indicator, only the maximum daily  1 hour average is assessed for the
34    PM mass  concentration indicator (below and Appendix A).  In looking  at the UFVA assessment
35    of PM light extinction, we noticed a close correspondence between the  level of protection
36    afforded for all 15 urban areas in the assessment by the maximum daily daylight 1-hour with 90th

                                                4-15
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1    percentile and the all daylight 1-hour with the 98th percentile (UFVA section 4.1.4). Thus, at
 2    these percentiles, either form of the standard provides protection of welfare effects to both the
 3    fraction of the public with rare opportunities and to those with ample opportunities to experience
 4    visibility. For consistency between PM2.5 mass and PM light extinction and because the daily
 5    maximum daylight 1-hour is thought to be more protective of those with limited opportunities to
 6    experience visibility, only the daily maximum form results will be displayed here.
 7    Temporal Variations of Visibility Impacts
 8           While visibility conditions can change quickly (i.e., less than a minute), atmospheric
 9    sight path averaged light extinction which is pertinent to visibility impacts generally changes
10    more slowly (i.e. tens of minutes generally).  Sub-hourly variations in light extinction
11    determined at any point in the atmosphere are likely the result of small scale spatial pollution
12    features (i.e. plumes) being transported by the wind across that point. At typical wind speeds
13    found in U.S. cities, an hour corresponds to a few tens of kilometers  of air flowing past a point,
14    which is similar to visibility sight path lengths of interest in urban areas.
15           PM concentrations and light extinction in urban environments vary from hour to hour
16    throughout the day due to a  combination of diurnal meteorological conditions and systematic
17    changes in emissions activity (e.g. rush hour traffic).  Generally low  mixing heights at night and
18    during the early morning hour tend to trap locally produced emissions, which are diluted as the
19    mixing height increases due to heating during the day.  Low temperatures and high relative
20    humidity at night are conducive to the presences of ammonium nitrate particles and water growth
21    by hygroscopic particles compared with the generally higher temperatures and lower relative
22    humidity later in the day. These combine to make early morning the most likely time for peak
23    urban visibility impacts.  Superimposed on such systematic time of day variations are the effects
24    of synoptic meteorology (i.e., those associated with changing weather) and regional scale air
25    quality that can generate peak visibility impacts any time of day.  The net effects of the
26    systematic urban and larger  scale variations are that peak daytime PM light extinction can  occur
27    any time of day though more often in early morning hours  (UFVA section 3.4.2 and 3.4.3 and
28    shown in Figures 3-9, 3-10 and 3-12). Use of multi-hour averaging of PM light extinction would
29    suppress peak hourly conditions that are expected to represent peak visibility impacts.
30    Meteorological Causes  of Visibility Impacts
31           Visibility is also reduced directly by the presence of precipitation and fog regardless of
32    the presence or absence of PM. A secondary PM NAAQS is not meant to protect against such
33    sources of visibility impairment not related to PM.  Therefore, one consideration in the
34    development of alternative standard forms (discussed below in section 4.3) was the inclusion of a
                                                4-16
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1    relative humidity screen in an effort to avoid misinterpretation of the direct effect of
 2    meteorological conditions on visibility and those caused by PM air quality.
 3    •  Based on currently available information, what range of levels of visibility impairment
 4       is reasonable to consider in reaching judgments about the adequacy of the current
 5       NAAQS?
 6          In order to identify levels of visibility impairment appropriate for consideration in setting
 7    secondary PM NAAQS to protect the public welfare, we comprehensively examined information
 8    that was available in this review regarding people's stated preferences for acceptable and
 9    unacceptable visual air quality.
10          Light extinction is an atmospheric property that by itself does not directly translate into a
11    public welfare effect. Instead, light extinction becomes meaningful in the context of the impact
12    of visibility on the human observer.  This has been studied in terms of the acceptability or
13    unacceptability expressed for it by a human observer.  The perception of the visibility impact of
14    a given level of light extinction occurs in conjunction with the associated characteristics  and
15    lighting conditions of the viewed scene.   Thus, a given level of light extinction may be perceived
16    differently by an observer looking at a different scene or the same  scene with different lighting
17    characteristics. Likewise, different observers looking at the same scene with the same lighting
18    may have different preferences regarding the associated VAQ. When scene and lighting
19    characteristics are held constant, the perceived appearance of a scene (i.e., how well the scenic
20    features can be seen and the amount of visible haze) depends only  on changes in light extinction.
21    This has been demonstrated using the WinHaze model that uses image processing technology to
22    apply user-specified changes in light extinction values to the same base photograph with set
23    scene and lighting characteristics.
24          Much of what we know about the acceptability of levels of visibility comes from survey
25    studies in which participants were asked questions about their preference or the value they place
26    on various visibility levels as displayed to them in scenic photographs and/or WinHaze images
27    with a range of known light extinction levels. Urban visibility preference studies for four urban
28    areas were reviewed in the UFVA (chapter 2) to assess the light extinction levels judged by the
29    participant to have acceptable visibility.   While the results differed among the four urban areas,
30    results from a rating exercise showed that within each preference study, survey participants
31    consistently distinguish between different levels of light extinction and prefer and value
32    visibility associated with lower light extinction levels among the scenic images they are shown.
33          The reanalysis included three completed urban visibility preference survey  studies plus a
34    pair of smaller focus studies designed to explore and further develop urban visibility survey
35    instruments. The three western studies included Denver, Colorado (Ely et al., 1991), one in the
36    lower Fraser River valley near Vancouver, British Columbia (BC), Canada (Pryor, 1996), and

                                                4-17
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1    one in Phoenix, Arizona (BBC Research & Consulting, 2003). A pilot focus group study was
 2    also conducted for Washington, DC (Abt Associates Inc., 2001).  In response to an EPA request
 3    for public comment on the Scope and Methods Plan (74 FR 11580, March 18, 2009), we
 4    received comments (Smith, 2009) about the results of a new Washington, DC focus group study
 5    that had been conducted using methods and approaches similar to the method and approach
 6    employed in the EPA pilot study (Smith and Howell, 2009). When taken together, these studies
 7    from the four different urban areas included a total of 852 individuals, with each individual
 8    responding to a series of questions answered while viewing a set of images of various urban
 9    VAQ conditions.
10           The approaches used in the four studies are similar and are all derived from the method
11    first developed for the Denver urban visibility study.  In particular, the studies all used a similar
12    group interview type of survey to investigate the level of visibility impairment that participants
13    described as "acceptable.  While each study asked the basic question, "What level of visibility
14    degradation is acceptable?", the term "acceptable" was not defined, so that each person's
15    response was based on his/her own values  and preferences for VAQ.  Given the similarities in
16    the approaches used, we concluded that it is reasonable to compare the results to identify overall
17    trends in the study findings and that this comparison can usefully inform the selection of a range
18    of levels for use in further analyses. However, variations in the specific materials and methods
19    used  in each study introduce uncertainties that should also be considered when interpreting the
20    results of these comparisons. Key  differences between the  studies include: 1) image presentation
21    methods (e.g., projected slides of actual photos, projected images generated using WinHaze  (a
22    significant technical advance in the method of presenting VAQ conditions), use of computer
23    monitor screen; 2) number of participants in each study; 3) participant representativeness of the
24    general population of the relevant metropolitan area; and 4) specific wording used to frame the
25    questions used in the group interview process.
26           In the UFVA,  each study was evaluated separately and figures developed to display the
27    percentage of participants that rated each photograph as "acceptable". The horizontal axis was in
28    terms of light extinction (deciview) and the vertical axis in  terms of percent of participants rating
29    "acceptable". Ely et al. (1991) introduced a "50% acceptability" criteria analysis of the Denver
30    preference study results. The 50% acceptability criteria is designed to identify the VAQ level
31    that best divides the photographs into two groups: those with  a VAQ rated as acceptable by the
32    majority of the participants, and those rated not acceptable by the majority of participants. We
33    adopted the criteria as a useful index for comparison between studies.  The results of each
34    individual analysis were then combined graphically to allow for visual comparison.  Figure 4-1
35    (Figure 2-16 in UFVA) presents the graphical summary of the results of the studies in the four
36    cities and draws on results previously presented in Figures 2-3, 2-5, 2-7 and 2-11 of chapter  2 in

                                                4-18
      Draft - Do Not Quote or Cite                                                   March 2010

-------
1
2
3
4
5
6
7
      the UFVA. Figure 4-1 also contains lines at 20 dv and 30 dv that effectively and pragmatically
      identify a range where the 50% acceptance criteria occur across all four of the urban preference
      studies. Out of the 114 data points shown in Figure 4-1, only one photograph (or image) with a
      VAQ below 20 dv was rated as acceptable by less than 50% of the participants who rated that
      photograph.2  Similarly, only one image with a VAQ above 30 dv was rated acceptable by more
      than 50% of the participants who viewed it.3

             Figure 4-1 Summary  of results of urban visibility studies in four cities, showing the
             identified range of the 50% acceptance criteria .4
                      2O Mm-1
                  100%  	:	*--=!«
SO Mm"1
                                                     1OO Mm-1  2OO Mm'1  4OO Mm *  8OO Mnrr1
A

Denver

*

Phoenix

#
BC
-BC Logit
* Washington
Hf* B ^nit

 9           As can be seen in the figure, each urban area has a separate and unique response curve
10    that appears to indicate that it is distinct from the others.  These curves are the result of a
11    logistical regression analysis using a logit model of the greater than 19,000 ratings of haze
12    images as acceptable or unacceptable. The model results can be used to estimate the VAQ
13    deciview values where the estimated response functions cross the 50% acceptability level, as
       Only 47% of the BC participants rated a 19.2 dv photograph as acceptable.
      3 In the 2001 Washington, D.C. study, a 30.9 dv image was used as a repeated slide. The first time it was shown
      56% of the participants rated it as acceptable, and 11% rated it as acceptable the second time it was shown. The
      same VAQ level was rated as acceptable by 42% of the participants in the 2009 study (Test 1).
      4 Top scale shows light extinction in inverse megameter units; bottom scale in deciviews. Logit analysis estimated
      response functions are shown as the color-coded curved lines for each of the four urban areas
                                                 4-19
     Draft - Do Not Quote or Cite
                                                                                    March 2010

-------
1
2
3
4
5
      well as any alternative criteria levels.  Selected examples of these are shown in Table 4-1 (Table
      2-4 in UFVA, Chapter 2). These results show that the logit model data also support the upper
      and lower ends of the range of 50th percentile acceptability values (e.g. near 20 dv for Denver
      and near 30 dv for Washington, DC) already identified in Figure 4-1 (see Table 2-4 UFVA).
 6    Table 4-1 Logit model estimated VAQ values corresponding to various percent
 7             acceptability values for the four cities.

90% Acceptability criteria
75% Acceptability criteria
50% Acceptability
Criteria
25% Acceptability criteria
10% Acceptability criteria
Denver
14.21
17.05
19.90
22.74
25.59
British
Columbia
16.80
19.63
22.45
25.28
28.10
Phoenix
24.15
21.80
24.15
26.51
28.87
Washington,
DC
23.03
26.03
29.03
32.03
35.03
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
           Based on the composite results and the effective range of 50th percentile acceptability
     across the four urban preference studies shown in Figure 4-1 and Table 4-1, benchmark levels
     have been selected in a range from 20 dv to 30 dv (74 Mm"1 to 201 Mm"1) for the purpose of
     provisionally assessing whether visibility conditions would be considered acceptable (i.e., less
     than the low end of the range), unacceptable (i.e., greater than the high end of the range) or
     potentially acceptability (within the range). A midpoint of 25  dv (122 Mm"1) was also selected
     for use in the assessment.  This level is also very near to the 50th percentile criteria value from
     the Phoenix study (i.e. 24.3 dv), which  is by far the best of the four studies in terms of least noisy
     preference results and the most representative selection of participants. Based on the currently
     available information, we conclude that the use  of 25 dv to represent the middle of the
     distribution of results seems well supported.
           These three benchmark values provide a low, middle, and high set of light extinction
     conditions that are used to provisionally define daylight hours with urban haze conditions that
     have been judged unacceptable by  the participants of these preference studies. As discussed
     above, PM light extinction is taken to be light extinction minus the Rayleigh scatter (i.e. light
     scattering by atmospheric gases which is on average about 10 Mm"1), so the low, middle and
     high levels correspond to PM light extinction  levels of about 64 Mm"1, 112 Mm"1 and 191 Mm"1.
     In the UFVA, these three levels were called Candidate Protection Levels (CPLs). We  continue
     to use this term in this document. However, it is important to note that the degree of protection
                                               4-20
     Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    provided by a secondary NAAQS is not determined solely by the level of the standard but by all
 2    the components (e.g., indicator, form, level, averaging time) being applied together. Therefore,
 3    the reader should keep in mind that the term CPL is meant only to indicate levels within a range
 4    that we feel are appropriate for consideration that could, in conjunction with other aspects of the
 5    standard, provide an appropriate degree of visibility protection.
 6          In characterizing our degree of confidence in each CPL and across the range, a number of
 7    issues were considered. Looking first at the two studies that define the upper and lower bounds
 8    of the range, we considered whether they represent a true regional distinction in preferences for
 9    urban visibility conditions between Western and Eastern U.S.. There is little information
10    available to help sort this out, especially given that we have preference studies in only one
11    Eastern urban area.  Smith  and Howell (2009) found little difference in preference response to
12    Washington, DC haze photographs between the study participants from Washington, DC and
13    those from Houston, TX. This provides some limited evidence that the value judgment of the
14    public in different areas of the country may not be an important factor in explaining the
15    differences in these study results.
16          In further considering what factors could explain the observed differences in preferences
17    across the four urban areas, we noted that the urban scenes used in each study had different
18    characteristics. For example, each of the Western urban visibility preference study scenes
19    included mountains in the background while the  single Eastern urban study did not. It's also true
20    that each of the Western scenes included objects at greater distances from the camera location
21    than in the Washington, DC study. There's no question that objects at a greater distance have a
22    greater sensitivity to perceived visibility changes as light extinction is changed compared to
23    otherwise similar scenes with objects at a shorter range.  This alone might explain the difference
24    between the results of the Washington DC, study and those from the Western urban studies.
25    Also it's intuitively likely that people value the views of mountains in the background more than
26    generic distant buildings in the foreground of the Western scenes; just as it seems obvious that
27    the Capital Mall and Washington Monument were the likely objects of greatest interest for the
28    Washington, DC study base photograph. Having scenes with the object of greatest intrinsic
29    value  nearer  and hence less sensitive for Washington compared with more distant objects of
30    greatest  intrinsic value in the Western urban areas could further explain the difference in
31    preference results.
32          Another question that we considered was whether the high CPL value that is based on the
33    Washington DC preference results is likely to be generally representative of urban areas that do
34    not have associated mountains or other valued objects visible in the distant background. Such
35    areas would include  the middle of the country and many areas in the eastern US. In order to
36    examine this issue, an effort would have to be made to see if scenes in such areas could be found

                                                4-21
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    that would be generally comparable to the western scenes (e.g., contain valued scenic elements at
 2    more sensitive distances than that used in the Washington, DC study). This is only one of a
 3    family of issues concerning how exposure to urban scenes of varying sensitivity affects public
 4    perception, for which no information is currently available. Additional urban visibility
 5    preference studies employing images selected of potentially more sensitive scenes could help
 6    evaluate whether a lower value for the high CPL is supportable.  Other investigations to
 7    determine how common such scenes are in various regions of the country would also be
 8    informative.  Until such information becomes available, the high end of the CPL range (30 dv)
 9    seems to be an appropriate level to consider.
10          With respect to the low end of the range, we considered factors that might further refine
11    our understanding of the robustness of this level. We concluded  that additional urban preference
12    studies, especially with a greater variety in types of scenes, including potentially more sensitive
13    Western urban scenes, could help evaluate whether a lower CPL  value than the currently selected
14    20 dv is supportable.  Further, the reason for the noisiness in data points around the curves
15    apparent in both the Denver and British Columbia  results compared to the smoother curve fit of
16    Phoenix study results could be explored.  One possible explanation that we identified is that
17    these older studies used photographs taken at different times of day and on different days to
18    capture the range of light extinction levels needed for the preference studies. By contrast, the use
19    of WinHaze in the Phoenix (and Washington,  DC) study, reduced variations in scene appearance
20    that affects preference rating and avoided the uncertainty inherent in using ambient
21    measurements to represent sight path averaged light extinction values. Reducing these sources
22    of noisiness and uncertainty in the results of future studies of sensitive urban scenes could
23    provide more certainty in the  selection of a low CPL value.
24    •   To what extent does the  available information demonstrate or suggest that PM-related
25       visibility impairment (within the range of CPLs) is occurring  at current ambient
26       conditions or at levels that would meet the current standards?
27    Current Visibility Levels
28          Chapter 3 of the UFVA characterized current visibility conditions in terms of both PM2.5
29    and light extinction levels for the  15 urban areas selected and compared them to the CPLs
30    identified above.
31          As an initial matter, we note that PM is not necessarily the primary source of visibility
32    impairment during periods with fog or precipitation.  In order to avoid precipitation and fog
33    confounding estimates of PM visibility impairment, and as advised by CASAC as part of its
34    comments on the first public review draft of the  UFVA, we restricted our assessment of visibility
35    conditions to daylight hours with relative humidity less than or equal to 90% (UFVA section
                                                4-22
     Draft - Do Not Quote or Cite                                                  March 2010

-------
 1    3.3.5 and Appendix G).  However, not all periods with relative humidity above 90% have fog or
 2    precipitation.  Removing those hours from application of a secondary PM standard involves a
 3    tradeoff between the benefits of avoiding many of the hours with meteorological causes of
 4    visibility impacts and the cost of not counting some hours with high relative humidity without
 5    fog or precipitation, where the growth of hygroscopic PM into large solution droplets results in
 6    enhanced PM visibility impacts. For the 11 urban areas included in the assessment for which
 7    updated meteorological data was obtained, a 90% relative humidity cutoff criterion is effective in
 8    that on average less than 6% of the hours are removed from consideration, yet those hours have
 9    on average twelve times the likelihood of meteorological causes of reduced visibility compared
10    with hours with 90% or lower relative humidity.
11           Figure 4-2 (Figure 3-8 in UFVA) presents box-and-whisker plots to  illustrate the
12    distributions of the estimates of daylight 1-hour reconstructed PM light extinction levels in each
13    area in each year during the 2005-2007 time period. The distribution of the daily maximum 1-
14    hour values is shown. The horizontal dashed lines in the plots represent the low, middle, and
15    high CPLs for PM light extinction of 64, 112, and 191 Mm"1, corresponding to the benchmark
16    VAQ values of 20 dv, 25 dv and 30 dv as discussed above.  Table 4-2 (Table 3-7 in UFVA)
17    provides the percentages of days (across all of 2005-2007, unweighted) in which the daily
18    maximum daylight 1-hour PM light extinction level was greater than each of the three CPLs
19    (excluding hours with relative humidity greater than 90 percent).
20           From these displays it can be seen that among the 15 urban areas, those in the East and in
21    California tend to have a higher frequency of visibility conditions above the high CPL compared
22    with those in the Western US. Both Figure 4-2 and Table 4-2 indicate that all 15 urban areas
23    have daily maximum hourly PM light extinctions that exceed even the highest CPL some of the
24    time. Again, the non-California western urban locations have the lowest frequency of maximum
25    hourly PM light extinction with values in excess of the high CPL for 8 percent or fewer of the
26    days. Except for the two Texas and the non-California western urban areas, all of the other
27    urban areas exceed that high CPL from about 20 percent to over 60 percent  of the days. Based
28    on these estimated maximum hourly PM light extinction estimates, all 15 of the urban areas
29    exceed the low CPL for about 40 percent to over 90 percent of the days. Based  on all of the
30    above, we conclude that current levels of PM light extinction associated with recent PM air
31    quality exceed levels that could reasonably be considered as protective of the public welfare.
32
33
34
                                               4-23
     Draft - Do Not Quote or Cite                                                  March 2010

-------
1    Figure 4-2. Distribution of estimated maximum daily daylight 1-hour PM light extinction

2               across the 2005-2007 period, by study area (excluding hours with relative

3               humidity greater than 90 percent). (Adapted from Figure 3-8 in UFVA)*


                                     Daily Maximum Extinction (Daylight Hours)
                            •V    Cr   ^       *•     ^   -0*   fS*   ,a~    ^   H--
                            />   Q^    Cj,   xrV    *Jr   x&    Krxs    ^.-    <«    .^9    A^    jy
                                ^    ^    v1^    \9    r?5    r^^   XiP    X^"    -<^    ^»^   ~^
                               V*    Cs^   ^-^   c^-    ^    ^^   ^3    -^   ^   -X^   ^
                          ^   c^"         ^        <$•              ^   **>    ^   ^



4           *In thel box-and-whisker plot, the box represents the 25th to 75th percentile range and the

5    whiskers represent the 10th and 90th percentile points of the data; individual data points below the 10th

6    percentile and above the 90fh percentile  are graphed as small circles.  The three dashed horizontal lines

1    represent the three CPL levels of 64, 112, and 191 Mm1.
                                                 4-24
     Draft - Do Not Quote or Cite
March 2010

-------
1
2
3
4
5
     Table 4-2  Percentage of daily  maximum hourly values of daylight PM light  extinction
                exceeding  CPLs (excluding  hours with  relative  humidity  greater than  90
                percent), (adapted from Table 3-7 in UFVA)
Study Area
Tacoma
Fresno
Los Angeles
Phoenix
Salt Lake City
Dallas
Houston
St. Louis
Birmingham
Atlanta
Detroit
Pittsburgh
Baltimore
Philadelphia
New York
Average
Number of Days with
Estimates
109
324
300
86
306
273
148
289
349
279
141
277
181
143
225
229
Candidate Protection Level
64 Mm -1
112 Mm -1
191 Mm -1
(a) Percentage of Daily Maximum Hourly
Values Exceeding CPL
52
75
90
42
44
80
79
98
89
91
87
85
80
86
83
77
22
52
83
7
17
41
45
78
65
75
68
57
50
64
59
52
4
30
62
1
8
10
11
40
34
31
43
26
23
31
28
26
 7   Visibility Levels That Just Meet Current Standards
 8          In the second draft of the UFVA, we modeled the "what if scenario based on simulating
 9   just meeting the current suite of PM2s secondary standards: 15 |ig/m3 annual average PM2 5
10   concentration and 35 |ig/m3 24-hour average PM2 5 concentration with a 98th percentile form,
11   averaged over three years. The steps needed to model the "what if conditions involve explicit
12   consideration of changes in PM2 5 components and are described here. First, we applied
13   proportional rollback to all the PM2.5 monitoring sites in each study area, taking into account
14   PRB PM2.5 mass, to "just meet" the NAAQS scenario for the area as a whole, not just at the
15   visibility assessment study site. The health risk assessment document (EPA 2010a) describes
16   this procedure in detail. The degree of rollback is controlled by the highest annual or 24-hour
17   design value, which in most study areas is from a site other than the site used in this visibility
18   assessment.  The relevant result from this analysis is the percentage reduction in non-PRB PM2 5
19   mass need to "just meet" the NAAQS scenario, for each study area. These percentage reductions
                                               4-25
     Draft - Do Not Quote or Cite
                                                                                March 2010

-------
 1   are shown in Table 4-4 of the UFVA.  Note that Phoenix and Salt Lake City meet the 15/35
 2   NAAQS scenario under current conditions, and require no reduction. PM2.5 levels in these two
 3   cities were not "rolled up." Second, for each day and hour for each PM2 5 component, we
 4   subtracted the PRB concentration from the current conditions concentration, to determine the
 5   non-PRB portion of the current conditions concentration.  Third, we applied the percentage
 6   reduction from step 1 to the non-PRB portion of each of the five PM2.5 components and added
 7   back the PRB portion of the component. Finally, we re-applied the IMPROVE algorithm, using
 8   the reduced PM2.5 component concentrations, the current conditions PMio-2.5 concentration for the
 9   day and hour, and relative humidity for the day and hour.  We then included the term for
10   Rayleigh scattering.
11          Figure 4-3 and Table 4-3 display the results of the rollback procedure as a box and
12   whisker plot of daily maximum daylight 1-hour PM light extinction and the percentage of daily
13   maximum hourly PM light extinction values exceeding the CPLs when just meeting the current
14   PM2.5 NAAQS scenario of 15/35 ug/m3 (excluding hours with relative humidity greater than 90
15   percent.  These displays show that at the current PM NAAQS level (i.e., 15/35) all of the eastern
16   urban areas and Los Angeles exceed the least restrictive CPL more than 10% of the time and that
17   only Tacoma would not exceed the least restrictive CPL more than 2% of the time.
                                              4-26
     Draft - Do Not Quote or Cite                                                 March 2010

-------
 2   Figure 4-3. Distribution of daylight 1-hour PM light extinction when rolled back to just
 3              meet current PM fine NAAQS across the 2005-2007 period, by study area
 4              (excluding hours with relative humidity greater than 90 percent). *

 5   NAAQS Scenario: 15 ug/m3 annual; 35 ug/m3 24-hour
 6   Displayed: Daily Max Daylight Light Extinction (excluding hours >90% RH)

                                        PMRollbackDailyMaxCase! NAAQS
         o
         o
         (M
         O
         O
         o
      1  8
      •—  oo
         o
      W  o
      *-  (O
         o
         o
         (M
                                                                    277
                                                                         1S1
                                                               f"iT-r  t   t
 l
 8           * In thel box-and-whisker plot, the box represents the 25th to 75th percentile range and the
 9   whiskers represent the ltfh and 90th percentile points of the data; individual data points below the ltfh
10   percentile and above the 90fh percentile are graphed as small circles.  The three dashed horizontal lines
11   represent the three CPL levels of 64, 112, and 191 Mm1
12
                                               4-27
     Draft - Do Not Quote or Cite
March 2010

-------
 1
 2
 3
 4
 5
 6
Table 4-3 Percentage of daily maximum hourly values of daylight PM light extinction
          exceeding CPLs when "just meeting" the current PMi.s NAAQS (15/35 ug/m3)
          (excluding hours with relative humidity greater than 90 percent). (Adapted from
          Table 4-7 in UFVA).
Study Area
Tacoma
Fresno
Los Angeles
Phoenix
Salt Lake City
Dallas
Houston
St. Louis
Birmingham
Atlanta
Detroit
Pittsburgh
Baltimore
Philadelphia
New York
Average
Number of Days with
Estimates
109
324
300
86
306
273
148
289
349
279
141
277
181
143
225
229
Candidate Protection Level
64 Mm -1
112 Mm -1
191 Mm -1
(a) Percentage of Daily Maximum Hourly
Values Exceeding CPL
43
54
85
44
24
81
75
97
84
90
80
78
78
85
76
72
10
30
69
6
9
41
41
74
55
71
61
48
48
61
45
45
1
10
39
1
4
10
11
36
24
25
33
16
19
28
19
18
 7
 8
 9
10
11
12
13
14
15
16
17
4.2.2   Summary
          In summary, we reach the preliminary conclusion that the available information in
this review, as described above and in the UFVA and ISA, clearly calls into question the
adequacy of the current suite of PM2 5 standards in the context of public welfare protection from
unacceptable  levels of visibility impairment, primarily in urban areas, and supports consideration
of alternative standards to provide appropriate protection.
       This preliminary conclusion is based first on the large percentage of days that exceed the
range of CPLs identified for consideration under both current and just meet PM air quality
conditions. In particular, under just meet conditions for the suite of secondary PM NAAQS (i.e.
15/35 ug/m3) greater than 10% of the days exceed the highest, least protective CPL of 191 Mm"1
                                         4-28
Draft - Do Not Quote or Cite
                                                                                 March 20JO

-------
 1    for 10 of the 15 urban areas.  When the middle CPL of 112 Mm"1 is considered,  12 of the 15
 2    cities have greater than 10% of their days exceeding with a range of 30 to 74% of days exceeding
 3    this level.  At the lowest CPL of 64 Mm"1, the percentage of days exceeding range from 24 to
 4    97% across the 15 urban areas.
 5           Second, we have reached the preliminary conclusion that the averaging times associated
 6    with the current suite of PM NAAQS are not well suited to protect against PM-related visibility
 7    impairment on the basis that short term exposures (e.g., 1 hour or less) is sufficient for an
 8    unacceptable level of visual air quality to be observed and the associated impacts of that
 9    observation to be registered by the observer.  Since some portion of the population may only
10    have the opportunity to observe one hour or less of ambient daylight visibility conditions, relying
11    on an averaging time as long as that of the current 24 hour and annual PM standards would make
12    it difficult to identify a requisite level of protection that would translate into appropriate
13    protection against the maximum daily value.  In addition, these longer averaging times also result
14    in the inclusion of nighttime conditions, for which the science needed to support identification of
15    appropriate levels of visibility protection is not well developed or understood.
16           In reaching the preliminary conclusion that the current suite of PM2.5 standards are
17    inadequate to provide the appropriate protection of the public welfare from known and/or
18    anticipated adverse effects by calling into question the adequacy of the current levels and
19    averaging times, it also seems reasonable to consider whether the current indicator of PM2 5
20    remains useful in relating ambient PM to its public welfare effect of visibility impairment.
21    Section 4.3 below discusses these and other considerations in its discussion of alternative
22    standards for consideration.

23    4.3   CONSIDERATION OF ALTERNATIVE STANDARDS
24           Having reached the conclusion that just meeting the current suite of PM2.5 standards
25    continues to allow levels of PM visual air quality that, based on the scientific evidence and
26    information available in this review, can reasonably be considered adverse to the public welfare,
27    this section will discuss alternative standards that could potentially provide requisite public
28    welfare protection from known and/or anticipated adverse effects. Any PM standard that results
29    in emissions reductions can be expected to result in visibility improvements, though meeting that
30    standard does not ensure an adequate or uniform degree of protection from adverse visibility
31    conditions.  A requisite level of public welfare protection can be facilitated by adopting a
32    standard specifically designed with indicator, averaging time and form that better mimics the
33    characteristics of the effect of interest.
                                                4-29
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    4.3.1  Nature of the Indicator
 2       •  To what extent does information provide support for considering a different
 3          pollution indicator(s) for PM to replace or supplement the PM2.s mass based
 4          indicator?
 5          EPA staff is considering two alternative indicators: PM2.5 mass and PM light extinction.
 6    PM2.5 mass is taken here to be the same indicator as is used by the current suite of PM NAAQS.
 7    PM light extinction is the contributions to light extinction by PMio under ambient conditions.
 8    EPA staff believes that the use of PM light extinction as an indicator is justified because it is a
 9    physically meaningful measure of the PM quantity that is most relevant and directly related to
10    visibility effects.  The basis for considering each indicator is discussed below.
11    PM Light Extinction Indicator
12          PM light extinction is highly related to light extinction, which is the property of the
13    atmosphere that is most directly related to visibility effects.  It differs from light extinction by the
14    nearly constant contributions for Rayleigh (or clean air) light scattering and the minor
15    contributions by NC>2 light absorption. Most importantly, there are no confounding issues
16    associated with particle size or composition, including PM water. PM light extinction can be
17    directly measured under ambient conditions by several instrumental methods, some of which
18    have been used for decades to routinely monitor the two components of PM light extinction
19    (light scattering and absorption) or to jointly measure both as total light extinction (from which
20    Rayleigh scattering is subtracted to get PM light extinction).
21          EPA staff also recognizes that while PM2.5 light extinction can be accurately measured by
22    a number of commercially available instruments, it is technically more challenging to measure
23    the PMio - 2.5 component of light extinction. This issue many not be of concern for many regions
24    of the country where the PMio -2.5 contributions to light extinction are minor (see UFVA section
25    3.4.5). However for some regions  (e.g. Phoenix), the PMio-2.5 contributions to light extinction
26    are substantial and may need to be  accounted for by some other methodology (e.g. estimation
27    from PMi0-2.5 mass concentration  measurements).  As an alternative to direct measurements,
28    light extinction can be estimated from PM speciation  and relative humidity data (see section 4.2).
29    There are a number of reasons for preferring direct measurements to algorithm-estimates of PM
30    light extinction for use in a secondary standard. These include the greater accuracy of direct
31    measurements, the ability to have short averaging times and the overall simplicity absent when
32    multiple measured parameters need to be processed to generate the indicator for a standard.
33          There currently is no FRM  for PM light extinction monitoring. Development of such an
34    FRM could take one to two years.  There is no routinely operated PM light extinction monitoring
35    program in urban areas. Deploying such a network will required time and resources. As a result,

                                                4-30
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    the selection of PM light extinction as an indicator could reasonably be expected to extend the
 2    time required for promulgation/implementation of a secondary PM NAAQS.
 3    PMi.g Mass Indicator
 4          PM2.5 mass is responsible for most of the visibility impairment in the majority of urban
 5    areas. In these areas, the contribution of PMio-2.5 is a minor contributor to visibility impairment
 6    most of the time. However, at some locations (see UFVA Figure 3-13 for Phoenix) PMio-2.5 is a
 7    major contributor to urban visibility effects. In those cities, a PM2.5 standard alone may not
 8    provide adequate visibility protection. In the absence of PM air quality information from a much
 9    larger number of urban areas across the country, it is not possible at this time to know how many
10    urban areas fall into this category, though it is reasonable to presume that other urban areas in the
11    desert southwestern region of the country may have conditions similar to the conditions shown
12    for Phoenix. A possible solution to this issue would be to add a companion PMio - 2.5 standard to
13    control under those circumstances. This latter refinement has not been developed for this review.
14          PM mass monitoring methods  are in wide spread use, including the Federal Reference
15    Method (FRM) involving the collection of periodic (1 day in 6 and 1 day in 3) 24-hour filter
16    samples. These samples are then subsequently analyzed to determine 24-hour PM fine mass.
17    The Chemical Speciation Network and continuous PM2.s monitoring produce hourly average
18    mass concentrations and are conducted at many locations. At a few locations, continuous
19    speciation  sampling produces measures some  of the major PM2.s chemical components. These
20    routine monitoring activities do not include measurement of the water content of the ambient PM
21    that contributes, often significantly, to visibility impacts.  Further, the PM mass concentration
22    monitoring does not provide information on the composition of the PM which also contributes to
23    the variability in the amount of visibility impact associated with any ambient PM mass
24    concentration.
25          The overall performance of PM2.s mass as a predictor of visibility effects as indicated by
26    PM light extinction can be seen in scatter plots shown in Figure 4.3-1 for Pittsburgh and
27    Philadelphia, PA (Similar plots for all 15 urban areas are in Appendix D, Figure D-2).  These
28    demonstrate the variation in hourly PM light extinction corresponding to any specific level of
29    PM2.5 mass concentration as well as statistical differences of the average relationships (depected
30    as the best fit lines) between cities. While this degree of variation is an important factor in
31    considering the appropriateness of a PM2.s mass indicator, we note that correctly identifying each
32    hour's visibility conditions is not required to estimate the extent to which the distribution of
33    visibility conditions may negatively impact public welfare.
                                                4-31
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    Figure 4.3-1 Scatter plots of PM light extinction versus PM2 s mass concentration of two
 2               cities (from UFVA Appendix D, Figure D2).
                  1000

               ~  800
                E
               —  600 -
                Q
               J=  400 -
                s
               ^  200 -
                                  Pittsburgh. PA
Philadelphia, PA
                                50
                                       100
                                               150
                                                      200 0
                                                                50
                                                                       100
                                                                               150
                                                                                      200
                                           PMi5 Mass Concentration
 4    4.3.2  Averaging and Applicable Times
 5          Consideration of an appropriate averaging time and factors with respect to the timing
 6    over which a secondary standard should apply were informed by consideration of the nature of
 7    PM visibility effects.
 8          Nighttime visibility impacts, described in the ISA (section 9.2.2) are significantly
 9    different from daytime impacts and not sufficiently well understood to be included at this time.
10    As a result, a secondary standard to protect visibility would best only apply to daylight hours.  In
11    the UFVA daylight hours were defined to be those morning hours having no minutes prior to
12    local sunrise and afternoon hours having no minutes after local sunset.  This definition ensures
13    the exclusion of periods of time where the sun is not the primary outdoor source of light to
14    illuminate scenic features.
15          PM is not necessarily the primary source of visibility impairment during periods with fog
16    or precipitation. By excluding daylight hours with average relative humidity above 90% for
17    inclusion in a secondary PM standard, the likely occurrence of visibility effects from fog and
18    precipitation is significantly reduced (UFVA section 3.3.5 and Appendix G). However not all
19    periods with relative humidity above 90% have  fog or precipitation. So removing those hours
20    from application of a secondary PM standard involves a tradeoff between the benefits if avoiding
21    many of the hours with meteorological causes of visibility impacts and not counting some hours
22    without fog or precipitation, but where the growth of hygroscopic PM to large solution droplets
23    results in enhanced PM visibility impacts.  For the 11 urban areas included in the assessment for
                                                4-32
      Draft - Do Not Quote or Cite
                 March 2010

-------
 1    which updated meteorological data was obtained,5 a 90% relative humidity cutoff criterion is
 2    effective in that on average less than 6% of the hours are removed from consideration, yet those
 3    hours have on average twelve times the likelihood of meteorological causes of reduced visibility
 4    compared with hours with 90% or lower relative humidity.
 5           As discussed above (section 4.2.1) selection of an appropriate averaging time takes into
 6    account both how quickly people experience and judge visibility conditions, as well as the
 7    typical rate of change of the path averaged PM light extinction over urban areas. While
 8    perception of visibility change can occur in less than a minute, meaningful changes to path
 9    averaged light extinction occur more slowly and can be well represented by hourly averaging.
10    Multi-hour averaging times would have the effect of reducing the magnitude of hourly peak
11    visibility impacts which can change significantly from one daylight hour to the next (see UFVA
12    Figure  3-12).  Reduction of peak values through multi-hour averaging reduces the ability of the
13    indicator to accurately characterize the visibility effects experienced by the segment of the
14    population that has infrequent short-term exposure  during peak periods.

15    4.3.3   Alternative Levels/Forms
16    Candidate Protection Levels
17           The results from the visibility preferences studies conducted in four urban areas define a
18    range of low, middle and high CPLs of 20 dv, 25 dv and 30 dv which are equivalent to PM light
19    extinction of values of 64 Mm"1, 112 Mm"1, and 191 Mm'^see section 4.2 above). With only the
20    four preference study results, the individual low and high CPL are in fact reflective of the results
21    from the Denver and Washington, DC studies in particular, and the middle CPL is very near to
22    the 50th percentile criteria result from Phoenix.
23           Determining PM2.s mass concentration values that correspond to the low, middle and
24    high CPL is complicated by the lack of a one-to-one relationship between PM light extinction
25    and PM2.5 mass. By considering a range of extinction efficiency values (i.e., the ratio of PM
26    light extinction to PM2.5 mass) from 3 m2/g to 10 m2/g, a range of PM2.5 concentration levels that
27    crudely corresponds to a low estimate of the low CPL and high estimate of the high CPL values
28    can be  determined. The full range is from ~6 ug/m3 to -60 ug/m3. Keeping in mind that there is
29    no exact one-to-one correspondence between PM mass concentration and PM light extinction, no
30    individual value in this range of conditions corresponds to the CPL values expressed in terms of
31    PM light extinction, so a number of alternate values within the range were selected for assessing
32    their visibility protection performance when combined with alternative forms (as described
      5 Through an oversight, EPA staff did not obtain NWS data for Los Angeles, St. Louis, Houston, and Detroit in time
      for processing. These data will be added in the final UFVA.
                                                4-33
      Draft - Do Not Quote or Cite                                                   March 2010

-------
 1    below).  Five alternative PM2.5 mass concentration values were selected for this purpose: 10
 2    ug/m3, 20 ug/m3, 30 ug/m3, 40 ug/m3 and 60 ug/m3.
 3    Alternative Forms
 4           The form of the current 24-hour PM2.5 NAAQS entails comparing the level of the
 5    standard to the three consecutive year average of the annual 98th percentile of the measured
 6    indicator.  The purpose in averaging for three years is to provide stability from the occasional
 7    effects of inter-annual meteorological variability that can result in unusual high pollution levels
 8    for a particular year that is otherwise typical. The use of a percentile form makes the standard
 9    less subject to the possibility of inappropriate violations caused by statistical outlier indicator
10    values. For consistency with the current PM NAAQS and to provide stability from unusual years
11    or outlier indicator data, the secondary PM standard should consider incorporating the use of a
12    three year average of a specified percentile.
13          The urban visibility preference studies that provided results leading to the range of CPLs
14    being considered in this document, offer no information that addresses the frequency of time that
15    visibility levels should be below those values.  Based on this and the nature of the public welfare
16    effect being one of aesthetics and/or feelings of wellbeing, we believe that it is not necessary or
17    appropriate to consider eliminating all such exposures and that allowing some number of
18    hours/days  with reduced visibility can reasonably be considered.  .  In the UFVA, 90th,  95th and
19    98th percentile annual values are included in the PM light extinction NAAQS scenarios  (Chapter
20    4). The hourly PM mass concentration scenarios that are described and assessed below and in
21    Appendix A include only the 90th and 95th percentile forms.6
22          Another aspect of the form that needs to be considered is whether to include all  daylight
23    hours or only the maximum daily daylight 1-hour. The maximum daily daylight 1-hour form is
24    more appropriate for protecting the welfare of people who have rare intermittent exposure to
25    visibility during the day (e.g. during commutes), but spend most of their time isolated from
26    outdoor views.  For such people a view of poor visibility during their morning  commute may
27    represent their perception of the day's visibility conditions until the next time they venture
28    outside during daylight, hours later or perhaps the next day.  Other people have exposure to
29    visibility conditions throughout the day.  For those people it might be more  effective to include
30    every daylight hour into assessing compliance with a standard, since a day with multiple hours
31    with visibility impairment is likely to be judged a greater impact on their wellbeing than a day
32    with just one  such hour.
      6 Assessment of the 98th percentile was thought to be unnecessarily restrictive, so it was not conducted at this time to
      simply and speed the assessment process.
                                                4-34
      Draft - Do Not  Quote or Cite                                                   March 2010

-------
 1           We do not know the fraction of the public that has only one or a few opportunities to
 2    experience visibility during the day, nor do we have information or studies on the duration of the
 3    effects on wellbeing associated with visibility conditions. However, it is logical that people with
 4    limited opportunities to experience visibility conditions on a daily basis would experience the
 5    entire impact associated with visibility based on their short term exposure. The impact of
 6    visibility for those who have access to visibility conditions often or continuously during the day
 7    may be based on varying conditions throughout the day. Based on these considerations, the
 8    segment of the population with infrequent access to visibility could be characterized as a
 9    susceptible population relative to peak visibility impairment, while those with longer exposures
10    are a susceptible population for longer-term visibility impairment.
11           In light of these considerations, the UFVA assessment of the various PM light extinction
12    scenarios included both forms and noticed a close correspondence between the level of
13    protection afforded for all 15 urban areas in the assessment by the maximum daily daylightl-
14    hour with 90th percentile and the all daylight 1-hour with 98th percentile (UFVA section 4.1.4).
15    In this sense, the reductions in visibility impairment required to meet either form of the standard
16    provides protection to both the fraction of the public with rare opportunities and to those with
17    ample opportunities to be affected by  PM-related visibility impairment.
18           Both forms are assessed for the PM light extinction indicator (UFVA), but only the
19    maximum daily form is assessed for the PM mass concentration indicator (below and Appendix
20    A).  For consistency between the two  indicators and because the daily maximum daylight 1-hour
21    is thought to be more protective of those with limited opportunities to experience visibility, only
22    the daily maximum form results will be displayed here.

23    4.3.4   Performance of Alternative Standards
24           We conducted assessments of alternative standards based both on hourly PM light
25    extinction as the indicator (see Chapter 4 of the UFVA) and on hourly PM mass concentration
26    indicator (see Appendix A in this document). There are nine PM light extinction alternative
27    standards discussed here (i.e.,  daily maximum daylight 1-hour PM light extinction at the low,
28    middle and high CPL for each of the 90th, 95th and 98th percentile forms) and there are ten PM2.5
29    mass concentration alternative standards (i.e., five concentration levels for each of the 90th and
30    95th percentile forms). In both cases the assessment involves rolling back non-policy relevant
31    background (PRB) PM light extinction or PM mass concentration values until these specific
32    alternative standards are just met.
33           In considering the performance of alternative standards, we focused on the uniformity
34    across the 15 urban areas of the resulting visibility  conditions, as measured in terms of light
35    extinction, when the alternative scenarios are just met. Because of the one-to-one

                                                4-35
      Draft - Do Not Quote or Cite                                                    March 2010

-------
 1    correspondence light extinction and PM light extinction,7 the PM light extinction indicator based
 2    standards should produce visibility conditions exactly as prescribed. The ability to structure a
 3    PM light extinction based alternative standard to the level of visibility protection desired is an
 4    advantage because it could affords the most uniform degree of visibility protection nationwide.
 5           Figures similar to 4.3-1 for the other PM light extinction scenarios are shown in
 6    Appendix F of the UFVA. Table 4.3-1 shows the design values for the 9 scenarios based on
 7    maximum daily 1-hour PM light extinction.  When an area just meets a scenario, its design value
 8    in principle should exactly equal the NAAQS level, so preparation of this table serves as a check
 9    against calculation errors. Note that the design values in Table 4.3-1, resulting from the rollback
10    steps described in section 4.1.4 of the UFVA, in some cases do not exactly equal the assumed
11    level of the NAAQS, although all  are quite close. In some cases (e.g. Phoenix for 191 Mm"V90th
12    and 95th percentile), current conditions already meet the scenario specifications so no rollback
13    was necessary and current design values are shown in Table 4.3-1 and reflected in box and
14    whisker plot figures. The minor differences between prescribed and assessed design values seen
15    for some applications of the rollback assessment are due to hours switching in the PM light
16    extinction frequency distribution that is purely an artifact of the rollback methodology as
17    described in the UFVA (section 4.3). These  discrepancies were judged too small to justify
18    iterative rollback that could have been used to eliminate them.
      7 In this assessment light extinction = PM light extinction + lOMm"1, where the last term is Rayleigh or clean air
      light scattering that in fact can range from about SMrn1 to HMm"1 depending on average sight path elevation above
      sea level.
                                                 4-36
      Draft - Do Not Quote or Cite                                                    March 2010

-------
1
2
3
4
5
Figure 4.3-2 Daily maximum daylight 1-hour light extinction after rollback to just meet a
           scenario with daily maximum of 112 Mm"1 for the 90th percentile excluding
           hours with relative humidity greater than 90 percent.
                              ExtRollbackDailyMaxNAAQS! 12Pctl90DVsFromdaily.max
 -C
 Dl
    o
    o
    CM
    O
    "
    O
    O
    o
    CO
    o
    o
    o
    o
                                    0


                                    0
                              /  ^
                                                                      ^    ^"
             ^   /
     - Do TVor  uote or Cite
                                           4-37
                                                                                   March 2010

-------
1
2
3
4
    Table 4.3-1. PM light extinction design values for "just meeting" secondary NAAQS
                scenarios based on measured PM light extinction (excluding hours with
                relative humidity greater than 90 percent)


Level (Mm ')
Percentile
Form

Tacoma, WA
Fresno, CA
Los Angeles,
CA
Phoenix, AZ
Salt Lake City,
UT
Dallas, TX
Houston, TX
St. Louis, IL
Birmingham,
AL
Atlanta, GA
Detroit, MI
Pittsburgh, PA
Baltimore, MD
Philadelphia,
PA
New York, NY
Secondary NAAQS Scenarios Based on Daily Maximum
(a)
191
90th
(b)
191
95th
(c)
191
98th
(d)
112
90th
(e)
112
95th
(f)
112
98th
(g)
64
90th
(h)
64
95th
(0
64
98th
PM light extinction Design Value
(based on same percentile form as the NAAQS scenario)
140
191
191
105
164
183
191
191
191
191
191
191
191
191
192
157
191
191
144
191
191
191
191
192
191
191
191
191
191
191
191
191
191
185
191
191
191
191
191
191
191
191
191
191
191
112
112
112
105
112
113
115
113
113
112
112
112
111
112
113
112
112
112
112
112
113
112
112
114
111
112
112
112
112
112
108
112
112
112
112
112
112
112
112
112
112
112
112
112
112
66
64
65
64
64
64
67
65
64
64
64
64
63
65
65
70
64
64
64
64
66
61
64
66
63
64
64
64
64
64
60
64
64
64
64
66
67
64
64
65
65
64
65
64
64
5
6
    Draft - Do Not Quote or Cite
                                            4-38
                                                                              March 2010

-------
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
            A complete description of the rollback assessment for the ten maximum daily 1-hour PM
     mass concentration scenarios is available in Appendix A of this document.  The process used is
     very similar to that used in the UF VA for rollback of the PM light extinction based scenarios.
            Since PM mass does not have a one-to-one correspondence to light extinction, the PM
     light extinction conditions resulting from just meeting standards based on PM mass are not as
     uniform as those shown above for PM light extinction scenarios.  This is demonstrated in the box
     and whisker plot of the maximum daily 1-hour PM light extinction for the 30 ug/m3, 90th
     percentile scenario shown below (Figure 4.3-2), which shows greater variation from among the
     urban areas than the PM light extinction based scenarios. Similar plots for all ten PM mass
     based scenarios are shown in Appendix A (Figure A-2).

     Figure 4.3-3 Maximum daily daylight 1-hour PM light extinction under "just meet"
                  conditions for a NAAQS scenario based on 1-hour PM2.s mass of 30 ug/m3,
                  90th percentile, excluding relative humidity >90%.
                                         PMRollbackDailyMaxNAAQS30Pctl90DVsFromdaily.max
                o
                o _
                o —
             5
             c
             o
             o
             '5
             LU
                O
                §
                                                             T   T  "-+  "-r "•*•"
                                                                                          T
                                                                 #•'
     Draft - Do Not Quote or Cite
                                               4-39
                                                                                  March 2010

-------
 1          The 90th percentile PM light extinction design values corresponding to the 90th percentile
 2   PM mass concentration based scenarios for the five mass concentration levels are shown in Table
 3   4.3-2. Values in the table that exceed each of the CPL values are highlighted using different
 4   colors.  The values in the table are the same for some urban areas for the less restrictive PM mass
 5   concentration standard (e.g. Tacoma at 60 ug/m3 and 40 ug/m3, or Phoenix at 60 ug/m3, 40
 6   ug/m3, 30 ug/m3 and 20 ug/m3), because those areas required no rollback to meet the less
 7   restrictive PM mass  concentration standards.  In order for most or all urban areas to achieve even
 8   the highest CPL at the 90th percentile, a PM2.5 mass based standard set below 40 ug/m3 would be
 9   necessary. The higher PM2.5 mass levels of 60 ug/m3 and 40 ug/m3 clearly are not sufficiently
10   protective, since they permit 11 and 10 of the 15  areas, respectively, to have design values larger
11   than 191 Mm"1. At the 20 ug/m3 PM2.5 standard level, all areas meet the highest CPL
12   benchmark, except for St. Louis, which doesn't attain the highest CPL even when a 10 ug/m3
13   standard is imposed.  We believe that this latter case may have occurred because the coarse mass
14   data in St. Louis may not be representative of the metropolitan area.8 The 30 ug/m3 standard is
15   marginally above the high CPL at three cities and only exceeds it substantially at St. Louis.
     8 The data set used to determine hourly PM mass and composition data for St. Louis may not be representative of the
     St. Louis metropolitan areas.  See the discussion in the UFVA section 3.2.1.
                                                4-40
     Draft - Do Not Quote or Cite                                                  March 2010

-------
 1
 2
 3
 4
 5
Table 4.3-2 90th percentile maximum daily 1-hour PM light extinction design values (Mm"
           *) after rollback to meet alternative standard of 60 ug/m3, 40 ug/m3, 30 ug/m3,20
           ug/m3 and 10 ug/m3 maximum daylight 1-hour PM mass concentration for the
           90th percentile.*

Tacoma, WA
Fresno, CA
Los Angeles, CA
Phoenix, AZ
Salt Lake City, UT
Dallas, TX
Houston, TX
St. Louis, IL
Birmingham, AL
Atlanta, GA
Detroit, MI
Pittsburgh, PA
Baltimore, MD
Philadelphia, PA
New York, NY
60 ug/m3
140 Mm'1
338 Mm'1
403 Mm'1
105 Mm'1
164 Mm'1
183 Mm'1
194 Mm'1
307 Mm'1
357 Mm'1
249 Mm'1
291 Mm'1
278 Mm'1
246 Mm'1
258 Mm'1
306 Mm'1
40 ug/m3
140 Mm'1
248 Mm'1
284 Mm'1
105 Mm'1
164 Mm'1
183 Mm'1
194 Mm'1
307 Mm'1
266 Mm'1
249 Mm'1
202 Mm'1
243 Mm'1
246 Mm'1
175 Mm'1
281 Mm'1
30 ug/m3
140 Mm'1
190 Mm'1
220 Mm'1
105 Mm'1
153 Mm'1
183 Mm'1
179 Mm'1
277 Mm'1
208 Mm'1
191 Mm"1
157 Mm'1
185 Mm'1
201 Mm'1
134 Mm'1
212 Mm'1
20 ug/m3
128 Mm'1
132 Mm'1
156 Mm'1
105 Mm'1
107 Mm'1
146 Mm'1
125 Mm'1
241 Mm'1
152 Mm'1
134 Mm'1
120 Mm'1
127 Mm'1
138 Mm'1
98 Mm'1
141 Mm"1
10 ug/m3
82 Mm'1
74 Mm'1
105 Mm'1
86 Mm'1
59 Mm'1
80 Mm'1
73 Mm'1
206 Mm'1
102 Mm'1
76 Mm'1
88 Mm'1
69 Mm'1
76 Mm'1
63 Mm'1
74 Mm'1
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
       * Colored highlighting shows which of the CPL levels the values are near, using the
following definitions: PM light extinction equal or above the high CPL >191 Mm'1; above the
middle CPL, 112 Mm1 -190 Mm1; above the low CPL, 64 Mm'1 - 111 Mm'1; below the low
CPL, < 64 Mm'1. Values without color highlighting indicate no rollback, so no information
about the effectiveness of the hourly PM mass based alternative standards

       Aside from St. Louis, the range of 90th percentile PM light extinction design levels
resulting from rollback to just meet a 90th percentile 20 ug/m3 PM mass concentration is from 98
Mm"1 for Philadelphia to 156 Mm"1 for Los Angeles with most values near the middle CPL value
of 112 Mm"1. A similar range, excluding St. Louis for the 30  ug/m3 PM mass concentration is
from 134 Mm"1 to 220 Mm"1 for the same two cities. This demonstrates the amount of variability
in visibility conditions that would likely result from using a standard with PM mass
concentration instead of PM light extinction as the indicator.  While this degree of variation is
not particularly large, it does mean that some areas would be required to further control
emissions to meet a secondary standard though they have visibility conditions that are as good or
better than other areas which meet the PM mass based standard.  Some measure of the extent to
                                               4-41
     Draft - Do Not Quote or Cite
                                                                            March 2010

-------
 1    which this occurs may be gained by comparing the percent rollback values required to meet the
 2    various PM mass and PM light extinction values.
 3          Based on inspection of the design values shown in Table 4.3-2, among the PM2 5 mass
 4    concentration levels that were assessed, 30 ug/m3 provides protection most comparable to the
 5    high CPL (191  Mm"1), 20 ug/m3 provides protection most comparable to the middle CPL (112
 6    Mm"1) and 10 ug/m3 provides protection most comparable to the low CPL (64 Mm"1).
 7          Care must be taken to avoid misinterpretation of this suggestion of approximate
 8    comparability of protection afforded the various urban areas by PM mass at the three
 9    concentration levels in terms of the three CPLs.  The values in Table 4.3-2 are the 90th percentile
10    values of the PM light extinction distribution that results from transforming the hourly PM mass
11    and composition values of the rolled back PM distributions for each urban area. Hours with
12    values above the 90th percentile of PM mass do not necessarily have PM light extinction above
13    the 90th percentile and visa versa.  As discussed earlier (section 4.3.1) for any individual hourly
14    PM mass concentration there is a substantial range of corresponding PM light extinction values
15    possible. However when used as the indicator for a secondary PM  standard, hourly PM2.s mass
16    need only predict the response of the distribution of hourly visibility conditions to produce
17    requisite levels of protection for visibility welfare effects.

18    4.4   PRELIMINARY STAFF CONCLUSIONS ON ALTERNATIVE SECONDARY PM
19         STANDARDS FOR VISIBILITY-RELATED EFFECTS
20              In reaching preliminary conclusions on potential alternative standards to provide
21    requisite protection of PM-related visibility impairment, staff has considered the basic elements
22    oftheNAAQS: indicator, averaging time, form and level. In considering the scientific and
23    technical information, we reflect upon the information available in the last review integrated with
24    information that is newly available since the last review as assessed and presented in the ISA and
25    the second draft UFVA (US EPA, 2010b) and as summarized in sections 4.2, 4.3, and Appendix
26    A.
27          As  outlined in section 4.1.3, in this review we emphasize a policy approach that broadens
28    the general approaches used in the last two PM NAAQS reviews by utilizing, to the extent
29    available, enhanced tools, methods, and data to more comprehensively characterize visibility
30    impacts. As such, we take into account both evidence-based and impact assessment-based
31    considerations to inform our conclusions related to the adequacy of the current PM2.5 secondary
32    standards and alternative standards that are appropriate for consideration in this review.  In so
33    doing, we are seeking to provide as broad an array of options as is supportable by the available
34    information, recognizing that the selection of a specific approach to reaching final decisions on
35    the secondary PM standards for protection from PM-related visibility impairment will reflect the

                                                4-42
      Draft - Do Not Quote or Cite                                                  March 2010

-------
 1   judgments of the Administrator as to what weight to place on the various approaches and types
 2   of information presented in the final PA.
 3           In presenting our preliminary conclusions regarding alternative secondary standards and
 4   ranges of levels for consideration, we summarize conclusions presented in sections 4.3.1 through
 5   4.3.4. We emphasize that these are preliminary conclusions that reflect consideration of the
 6   scientific and technical information assessed and presented in the ISA and second draft UFVA
 7   (US EPA, 2009a; US EPA, 201 Ob). We note that staff conclusions to be presented in the final
 8   UFVA and second draft PA will be based, in part, on input received from CASAC and the public
 9   on the second draft UFVAA and this first draft PA.
10           We recognize that selecting from among alternative standards will necessarily reflect
11   consideration of the qualitative and quantitative uncertainties inherent in the relevant evidence
12   and in the assumptions that underlie the quantitative visibility impact assessment. In identifying
13   these alternative secondary standards and ranges of levels for consideration, we are mindful that
14   the Clean Air Act requires standards to be set that are requisite to protect public welfare from
15   any known or anticipated adverse effects, such that the standards are to be neither more nor less
16   stringent than necessary. Thus, the Act does not require that the NAAQS be set at zero-risk
17   levels, but rather at levels that avoid unacceptable risks to public welfare.
18
19       ( 1)Consideration should be given to revising the current suite  of PM2.5 secondary standards
20           to provide increased public welfare protection from PM-related visibility impairment,
21           primarily in urban areas.  This preliminary conclusion is based in general on the
22           evaluation in the ISA of the currently available information, including a more extensive
23           characterization of the sources contributing to visibility  impairment in both rural  and
24           urban locations, a refined understanding of the contributions of various PM components
25           in such areas,  exposure-related evidence supporting a causal relationship between
26           ambient PM and impaired VAQ, and more specifically,  on  the evidence that a significant
27           number of days with levels of VAQ that could reasonably be considered unacceptable
28           based on the preference studies would continue to occur in  areas where the current
29           standards were met, together with judgments as to the public welfare significance of these
30           occurrences upon just meeting the current suite of PM2.5 standards.
31
32       ( 2) With regard to indicator,  consideration should be given to establishing a new PM light
33           extinction indicator; alternatively, consideration can be given to  retaining the current
34           PM2.5 indicator.  This preliminary conclusion takes into consideration the available
35           evidence that demonstrates a one-to-one correspondence between measured ambient PM
36           light  extinction and PM-related visibility impairment as well  as the significant degree of
37           variability in visibility protection across the U.S. allowed by a PM2.5 indicator, while also

                                                4-43
     Draft - Do Not Quote or Cite                                                   March 2010

-------
 1           recognizing that either indicator could be used as a basis for a standard that could provide
 2           appropriate protection from PM-related visibility impairment.
 3
 4       ( 3)With regard to averaging time, consideration should be given to selecting an alternative
 5           averaging time for a secondary PM standard to protect against daytime PM-related
 6           visibility impairment that takes into account the short term (instantaneous) nature of the
 7           perception of visibility impairment, short term variability in PM-related VAQ (partial
 8           hour to hourly), and the short-term nature of relevant exposure periods for the viewing
 9           public (partial hour to multiple hours).  Recognizing that the current 24-hour and annual
10           averaging times do not appropriately reflect these exposure characteristics, consideration
11           should be given to a one-hour averaging time based on the maximum hour in the daylight
12           period or on all daylight hours.
13
14       ( 4)In conjunction with considering a 1-hour averaging time, consideration should be given
15           to a form defined in terms of the 90th, 95th or 98th percentile of the distribution of 1-hour
16           PM light extinction or PM2 5 concentrations, averaged over three years.  We believe it is
17           appropriate to consider allowing some number of days/hours with visibility impairment
18           above the target level of protection, such that consideration of the 90th or 95th percentile
19           forms is appropriate. In addition, consideration should be given to applying a 90%
20           relative humidity screen to remove hours in which fog or precipitation is much more
21           likely to contribute to the observed visibility impairment.
22
23       ( 5) Consideration should be given to selecting a target level in terms of PM light extinction.
24           Further, consideration should be given to alternative candidate levels, with a particular
25           focus on a level of 112 Mm"1 as well as levels down to 64 Mm"1 and up to 191 Mm"1 to
26           provide appropriate protection against PM-related visibility impairment. Based on the
27           assessment conducted  in this review, we judge that a reasonably consistent
28           correspondence exists  between these PM light extinction benchmark levels and PM2 5
29           mass concentrations, with a particular focus on a level of 10 ug/m3 as well as levels down
30           to 20 ug/m3 and up to 30 ug/m3', which would be appropriate to consider in conjunction
31           with consideration of a PM2.5 indicator.
32
33       ( 6) Consideration should be given to the following alternative secondary PM standards to
34           provide protection against PM-related visibility impairment during daylight hours:
35
36           (a) Consideration of a new 1-hour daily maximum PM light extinction standard set at a
37              level within the range of 64 to 191 Mm"1 (e.g. 20 - 30 deciviews (dv)) with a 90th or
38              95th percentile form.
39

                                                4-44
      Draft - Do Not Quote or Cite                                                   March 2010

-------
  1            (b) Consideration of a revised 1-hour daily maximum PM2.5 standard set at a level within
  2                the range of 10 to 30 |ig/m3 with a 90th or 95th percentile form.
  3
  4            (c) Alternatively, consideration could also be given to a standard based  on all daylight
  5                hours with either indicator, in conjunction with consideration of a 98th percentile
  6                form.
  7
 9    4.5    REFERENCES

10    ABT (2001). Assessing Public Opinions on Visibility Impairment Due to Air Pollution: Summary Report.

11    BBC Research & Consulting. (2003), Phoenix Area Visibility Survey. Draft Report.. Available:
12            http://www.azdeq.gov/environ/air/download/vis 021903f.pdf. Accessed 9/16/2008

13    Henderson, R. (2005a). EPA's Review of the National Ambient Air Quality Standards for Paniculate Matter
14            (Second Draft PM Staff Paper, January 2005): A review by the Paniculate Matter Review Panel of the EPA
15            Clean Air Scientific Advisory Committee.  June 6, 2005. Available: http://www.epa.gov/sab/pdf/casac-05-
16            007.pdf.

17    Henderson, R.; Cowling, E.; Crapo, J.D..; Miller, F. J.; Poirot, R.L.; Speizer, F.; Zielinska, B. (2006b). Letter from
18            Clean Air Scientific Advisory Committee to the Honorable Stephen L. Johnson, Administrator, US EPA.
19            Clean Air Scientific Advisory Committee Recommendations Concerning the Final National Ambient Air
20            Quality Standards for Paniculate Matter. September 29, 2006.  Available:
21            http://www.epa.gov/sab/pdf/casac-ltr-06-002.pdf

22    Ely DW; Leary JT; Stewart TR; Ross DM (1991). The establishment of the Denver Visibility Standard. Presented at
23            Presented at: 84th annual meeting & exhibition of the Air & Waste Management Association; June;
24            Vancouver, British Columbia. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-48.4.

25    Malm, W.C.,  Sisler, J.F., Huffman, D., Eldred, R. A., Cahill, T. A. (1994) Spatial and Seasonal Trends in Particle
26            Concentration and Optical Extinction in the United States, Journal of Geophysical Research (Atmospheres)
27            99:1347-1370.

28    Pryor SC (1996). Assessing public perception of visibility for standard setting exercises. Atmos Environ, 30: 2705-
29            2716.

30    Smith, A. (2009). Comments on the First External Review Draft of EPA's "Integrated Science Assessment for
31            Paniculate Matter." CRA Internation. Washington, DC. March 30. Prepared for the Utility Air Regulatory
32            Group,. Submitted as public comments to the public meeting EPA Clean Air Science Advisory Council.
33            April 2.

34    Smith AE; Howell S (2009). An assessment of the robustness of visual air quality preference study results. CRA
3 5            International. Washington,
36            DC.http://yosemite.epa.gov/sab/sabproduct.nsf/B55911DF9796E5E385257592006FB737/$File/CRA+VA
3 7            Q+Pref+Robustness+Study+3+30+09+final.pdf.

38    US EPA (2005). Review of the National Ambient Air Quality Standards for Paniculate Matter: Policy Assessment
39            of Scientific and Technical Information, O AQPS Staff Paper. Research Triangle Park, NC 27711: Office
40            of Air Quality Planning and Standards; report no. EPA EPA-452/R-05-005a. December 2005. Available:
41            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm cr  sp.html

                                                      4-45
      Draft - Do  Not Quote or Cite                                                          March 2010

-------
  1    US EPA (2008a). Integrated Review Plan for the National Ambient Air Quality Standards for Paniculate Matter.
  2             National Center for Environmental Assessment and Office of Air Quality Planning and Standards, U.S.
  3             Environmental Protection Agency, Research Triangle Park, NC. Report No. EPA 452/R-08-004. March
  4             2008. Available at: http://www.epa.gov/ttn/naaqs/standards/pm/s_pm 2007_pd.html.

  5    US EPA (2009a). Integrated Science Assessment for Paniculate Matter:  Final Report. National Center for
  6             Environmental Assessment-RTF Division, Office of Air Quality Planning and Standards, Research
  7             Triangle Park, NC. EPA/600/R-08/139F.  December  2009.  Available:
  8             http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_2007 isa.html.

  9    US EPA (2009b). Integrated Science Assessment for Paniculate Matter: Second External Review Draft.  National
10             Center for Environmental Assessment-RTF Division,  Office of Air Quality Planning and Standards,
11             Research Triangle Park, NC.  EPA/600/R-08/139B. July 2009. Available:
12             http://www.epa.gov/ttn/naaqs/standards/pm/s_pm_2007 isa.html.

13    US EPA (2009d). Paniculate Matter National Ambient Air Quality Standards:  Scope and Methods Plan for Urban
14             Visibility Impact Assessment. Office of Air Quality Planning and Standards, U.S. Environmental
15             Protection Agency, Research  Triangle Park, NC. EPA-452/P-09-001. February 2009. Available:
16             http://www.epa.gov/ttn/naaqs/standards/pm/s_pm 2007_pd.html.

17    US EPA (20091). Paniculate Matter Urban-Focused Visibility Assessment - External Review Draft.  Office of Air
18             Quality  Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC.
19             EPA-452/P-09-005. September 2009.  Available:
20             http://www.epa.gov/ttn/naaqs/standards/pm/sjm 2007 risk.html.

21    US EPA (2010b). Paniculate Matter Urban-Focused Visibility Assessment - Second External Review Draft.  Office
22             of Air Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park,
23             NC. EPA-452/P-10-002. January 2010. Available:
24             http://www.epa.gOv/ttn/naaqs/standards/pm/s pm 2007 risk.html.
                                                       4-46
      Draft - Do Not Quote or Cite                                                          March 2010

-------
 i                              Appendix A

 2
 3    Information Regarding the 1-hour PM2.s Mass Indicator
 4
 5       This Appendix presents information on 2005-2007 levels of 1-hour PM2.5 mass
 6    concentrations in the 15 urban study areas and on the "what if PM light extinction conditions
 7    that would exist if the study areas met each of 10 alternative secondary PM NAAQS scenarios
 8    based on a 1-hour PM2.5 mass indicator.  With respect to the latter subject, this Appendix is
 9    therefore similar to Chapter 4 of the Particulate Matter Urban-Focused Visibility Assessment
10    (UFVA), which presented similar information for 18 secondary PM NAAQS scenarios based on
11    PM light extinction as the indicator, for the current annual and 24-hour PM2.5 NAAQS, and for a
12    scenario with an annual NAAQS of 12 |ig/m3 and a 24-hour NAAQS of 25 |ig/m3.
13
14    1.0 Indicator and Monitoring Method
15
16       As in Chapter 4 of the PM UFVA, this Appendix excludes from all NAAQS scenarios  and
17    results all non-daylight hours and all daylight hours with relative humidity greater than 90
18    percent.  This applies to both the definition of 10 secondary NAAQS scenarios, and to graphics
19    and tables that characterize ambient conditions.  While ambient humidity should not affect
20    conventional measurement approaches for 1-hour PM2.5 mass, the issue of co-occurrence of high
21    humidity levels with light extinction due to natural conditions would still apply. See section 3.3.5
22    of the UFVA.  The assumed hours of daylight are the same as those used in the UFVA, as  shown
23    in Table 3-5 of the UFVA.
24
25       All values for 1-hour PM2.5 mass concentration in this appendix come from the continuous
26    instruments at the 15 urban study sites, with no adjustment to make these values consistent with
27    the collocated 24-hour FRM measurement of PM2.5 mass. Appendix A of the UFVA provides
28    details on the type of continuous instrument at each study site.  TEOMs were used at all  sites
29    except for beta attenuation instruments in Fresno and Philadelphia, nephelometer instruments in
30    Tacoma and Phoenix, and an FDMS instrument in  Salt Lake City.
31
32       For conciseness in this first public review draft, only the daily maximum daylight 1-hour
33    PM2.5 mass concentration indicator is considered in this Appendix. It would also  be possible to
34    construct alternative NAAQS scenarios of an all-hours type, which could be analyzed in the
35    same manner as presented in this Appendix.
36
37    2.0 Current Conditions of 1-hour PMi.s Mass
38
39    Figure A-l is a box plot of 2005-2007 daily maximum daylight 1-hour PM2.5 mass
40    concentrations for the 15 study  areas, excluding hours with relative humidity greater than 90
41    percent, to give a sense of the range and central tendency of this parameter. The horizontal
42    reference lines are at 10,  20, 30, 40 and 60 |ig/m3. The relative positions of the 90 percentile
43    concentrations (indicated by the horizontal stroke at the top of the whisker) are generally
44    consistent with the relative ranking of these sites according to their design values  for the 24-hour

      Draft - Do Not Quote or Cite               A-1                              March 2010

-------
1
2
3
4
5
6
7
      PM2.5 NAAQS (see Table 3-2 of the UFVA); similarly, the relative positions of the median
      concentrations are generally consistent with the annual PM2 5 design values. Table A-l, based on
      the same data as Figure A-l, presents the percentage of days in 2005-2007 on which the daily
      maximum daylight 1-hour PM2.5 concentration exceeded the reference levels represented by the
      horizontal lines in Figure A-l.

      Figure A-l.  2005-2007 daily maximum daylight 1-hour PM2.s mass concentrations (ug/m3)
      for the 15 study areas (excluding hours with relative humidity greater than 90 percent)
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
                 109   324   300

                           O
                                    306   273   148   289   349    279    141   277   181    143    225
                -r

                                                                     of5    3*  ^T


     Draft - Do Not Quote or Cite
                                               A-2
March 2010

-------
1   Table A-l. Percentage of days with daily maximum daylight 1-hour PM2.s mass
2   concentration exceeding reference levels in 2005-2007 (excluding hour with relative
3   humidity greater than 90 percent)

Study Area
Tacoma
Fresno
Los Angeles
Phoenix
Salt Lake City
Dallas
Houston
St. Louis
Birmingham
Atlanta
Detroit
Pittsburgh
Baltimore
Philadelphia
New York

Number of Days
with Estimates
109
324
300
86
306
273
148
289
349
279
141
277
181
143
225
1-hour PM2 5 Mass Reference Level (ug/m3)
10
50
88
92
60
64
75
80
93
92
86
92
94
90
99
75
20
11
62
81
8
20
25
42
47
60
56
72
57
46
84
43
30
1
37
67
1
11
5
14
18
37
28
52
28
22
63
25
40
0
20
46
1
7
0
5
7
23
10
36
15
8
45
13
60
0
8
20
1
2
0
0
2
8
1
13
o
J
1
20
o
5
    Draft - Do Not Quote or Cite
A-3
March 2010

-------
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
         To ensure examination of a wide enough range of alternative standards based on 1-hour
     PM2.s mass to encompass the range of standards that might be considered as alternatives to the
     PM light extinction NAAQS scenarios examined in Chapter 4 of the UFVA, we considered
     levels of 10, 20, 30, 40, and 60 |ig/m3. Only the daily maximum daylight hour form was
     considered.  Each level was combined with two statistical forms: the three-year average of the
     annual 90th percentile value and the three-year average of the annual 95th percentile value. For
     ease of reference, these scenarios are designated by letters from "aa" to "jj" and listed in Table
     A-2. Looking somewhat ahead to results presented below, the scenarios are arranged in Table
     A-2 in order of least to most stringent in terms of the reductions in ambient PM2.5 needed from
     current levels to meet the current and alternative NAAQS levels and forms.

     Table A-2. Alternative NAAQS scenarios based on daily maximum daylight 1-hour PM2.s
     mass, averaged over three years (excluding hours with relative humidity greater than 90
     percent)
NAAQS Scenario
aa
bb
cc
dd
ee
ff
gg
hh
ii
jj
Level (ug/m3)
60
60
40
40
30
30
20
20
10
10
Statistical Form
3 -year average of 90th percentile
3-year average of 95th percentile
3 -year average of 90th percentile
3-year average of 95th percentile
3 -year average of 90th percentile
3 -year average of 95th percentile
3 -year average of 90th percentile
3-year average of 95th percentile
3 -year average of 90th percentile
3 -year average of 95th percentile
4.0 Approach to Modeling "What If Conditions of PM Light Extinction for Alternative
Secondary NAAQS Based on 1-hour PM2.5 Mass

   Before modeling "what if conditions, we augmented the data set described in Table 4 of the
UFVA in the same manner as described in Section 4.1.4 of the UFVA, to achieve seasonal
balance despite the lack of monitoring data for one quarter in each of Houston and Phoenix. In
Tacoma and Phoenix, which had data only for two years in the 2005-2007 period, we averaged
the percentile values from the only two available years rather than the three years defined for the
statistical form of the NAAQS scenarios.

   The modeling of daily maximum daylight 1-hour PM2 5 mass under each of the scenarios
listed in Table A-2 used a rollback approach that combined relevant concepts and steps from the
rollback methods described in sections 4.1.4 (for PM light extinction scenarios) and 4.2.2 (for
scenarios based on annual average and 24-hour average PM2.5) of the UFVA. The following are
the steps in the modeling.

1.  Identify the 90th percentile daily maximum daylight 1-hour PM2.5 mass value in each of 2005,
2006, and 2007 for a study area. Average these to determine the design value for that percentile
     Draft - Do Not Quote or Cite
                                         A-4
March 2010

-------
 1    form.  Repeat for the 95th percentile form. These design values are presented in Table A-3.
 2    They range from 22 to 81 |ig/m3 indicating that some study areas meet some of the NAAQS
 3    scenarios under current conditions. In such cases, PM2.5 concentrations were not adjusted, i.e.,
 4    there was no "roll up" for any area in any scenario.
 5
 6    2.  Using the same days and hours as contributed by the three annual 90th percentile values for
 7    actual 1-hour PM2.5 mass, find the three corresponding values of policy relevant background
 8    (PRB) 1-hour PM2 5 mass. Average these three annual values of PRB 1-hour PM2 5 to obtain the
 9    PRB portion of the actual 1-hour PM2.5 design value for the 90th percentile form.  Repeat for the
10    95th percentile form.
11
12       In the modeling for the NAAQS scenarios examined in the UFVA, PRB for 1-hour PM2 5
13    mass was not explicitly calculated because it was not needed in the rollback modeling for the
14    scenarios addressed in the UFVA. Therefore, it was necessary to reconstruct this parameter by
15    adding the  values for the PRB concentrations of the five components of PM2 5:  nitrate, sulfate,
16    elemental carbon, organic carbon material, and soil. The method for estimating PRB for these
17    five components is described in Appendix C of the UFVA.9
18
19    3.  Subtract the value from step 2 from the value from step 1, to determine the  non-PRB portion
20    of the 1-hour PM2.5 mass design value.
21
22    4.  Calculate the percentage reduction required in non-PRB 1-hour PM2 5 mass in order to reduce
23    the design value to the level that defines the NAAQS scenario, using the following equation:
24
25     Percent reduction required = 1 - (NAAQS level - PRB portion of the design value)/(non-PRB
26                                   portion of the design value)
27
28       The percentage reductions determined in step 4 are shown in Table A-4. Note that for some
29    combinations of area and scenario no reduction is required because the 2005-2007 design value
30    already meets the NAAQS scenario.
31
32    5.  Turning to the entire set of day/hour-specific actual and PRB daylight 1-hour concentrations
33    of the five PM2 5 components for the three (or two) year period, determine the non-PRB portion
34    of each of the five components in an hour by subtracting the PRB value from actual value,
35    reduce it by the percentage determined in step 4, and add back in the PRB 1-hour concentration
36    of the component.
37
38    6.  Finally, re-construct PM light extinction using the reduced values of the five components, the
39    original value of 1-hour PMio-2.s, and the 1-hour value of f(RH), according to the following
40    equation for PM light extinction (see section 3.2.3 of the UFVA for an explanation of the
41    variables in this equation).
42
      9 Table C-l of the UFVA presents the annual average of all daylight hour PRB PM2 5 mass. The 90th and 95th
      percentile values of daily maximum PRB PM2 5 mass were of course higher than shown in that table.

      Draft - Do Not Quote or Cite               A-5                              March 2010

-------
           = 3 x f(RH) x [Sulfate]
 2    + 3 xf(RH)x [Nitrate]
 3    + 4 x [Organic Mass]
 4    +10 x [Elemental Carbon]
 5    + 1 x [Fine Soil]
 6    + 0.6 x [Coarse Mass]
 7
 8       These steps assume that in order to meet a PM NAAQS scenario based on 1-hour PM2.s as
 9    the indicator, each component of PM2.5 is reduced by an equal percentage, across the five
10    components and across all hours. In actual implementation of such a NAAQS, each state would
11    develop an attainment strategy, which might result in unequal percentage reductions of the
12    components.  If the strategy emphasized reductions in the fine soil component, for example, PM
13    light extinction levels would remain high relative to those estimated by these steps, because fine
14    soil  is not efficient in terms of reducing visibility compared to the other four components on a
15    dry mass-to-mass basis.  On the other hand, a strategy that involves relatively large reductions in
16    sulfate or nitrate would achieve greater reductions in PM light extinction than estimated by these
17    steps.  The uncertainty in how the results of this rollback method compare to the results of actual
18    attainment strategies should be kept in mind when comparing the results of "what if scenarios
19    for NAAQS based on PM2.5 mass as the indicator versus scenarios based on PM light extinction.
20    Unlike the effect of humidity variation between areas, this source of uncertainty is not reflected
21    in any of the results presented in this Appendix and will not be apparent in comparisons of
22    results in this Appendix to results presented in the PM UFVA for NAAQS scenarios based on
23    PM  light extinction.
24
25       These steps also assume no change in PMio-2.5 concentrations between current conditions  and
26    "what if conditions. While reductions in PMio-2.s would not be needed to meet a secondary
27    NAAQS based on 1-hour PM2.5 mass, it is possible that strategies to control PM2.5 concentrations
28    might also achieve reductions in PMio-2.5 concentrations because some source emit both and
29    some control methods achieve some reductions in both. However, in most of the 15 study areas,
30    PMio-2.5 makes a small contribution to estimated PM light extinction, in part because in many of
31    the areas no local data on PMio-2.5 concentrations were available and the method used to fill this
32    gap  (application of a factor to PM2.s concentration) simply could not produce a high estimate  of
33    PMio-2.5.
     Draft - Do Not Quote or Cite                A-6                               March 2010

-------
1
2
Table A-3. 2005-2007 design values for 1-hour PM2.5 mass (ug/nr )
4
5
6
7
Study Area

Tacoma
Fresno
Los Angeles
Phoenix
Salt Lake
City
Dallas
Houston
St. Louis
Birmingham
Atlanta
Detroit
Pittsburgh
Baltimore
Philadelphia
New York
Percentile Form
90th
22
55
72
20
32
26
33
36
55
40
64
46
37
67
44
95th
27
66
81
24
45
29
37
44
74
45
79
51
43
77
55
Table A-4. Percentage reductions in non-PRB PM2.s components required to meet NAAQS
scenarios based on 1-hour PM2.s mass
Scenario
Level (ug/m3 )
Percentile Form
Study Area
Tacoma
Fresno
Los Angeles
Phoenix
Salt Lake City
Dallas
Houston
St. Louis
Birmingham
Atlanta
Detroit
Pittsburgh
Baltimore
Philadelphia
New York
aa
60
90
bb
60
95
cc
40
90
dd
40
95
ee
30
90
ff
30
95
gg
20
90
hh
20
95
ii
10
90
.ii
10
95
Percentage Reduction
0
0
17
0
0
0
0
0
0
0
7
0
0
10
0
0
10
26
0
0
0
0
0
19
0
24
0
0
22
0
0
28
45
0
0
0
0
0
28
0
38
13
0
40
8
0
40
51
0
12
0
0
10
46
12
50
22
8
49
28
0
46
59
0
7
0
9
16
46
25
54
35
19
55
32
0
55
64
0
34
0
20
33
60
34
63
42
31
62
46
11
65
73
0
39
23
40
45
65
51
70
57
47
71
55
27
71
76
15
56
34
49
56
74
57
75
62
55
75
65
60
83
87
51
70
64
71
74
84
77
85
79
74
86
78
69
86
88
58
78
69
78
79
87
80
88
81
78
88
83
    Draft - Do Not Quote or Cite
                                       A-7
March 2010

-------
 1    5.0. 1-hour PM2.s Mass Results for "Just Meeting" Alternative Secondary NAAQS
 2    Scenarios Based on 1-hour PM2.s Mass
 O
 4       As a check on the reasonableness of the rollback method described in section 4.0 and on the
 5    accuracy of the code used to implement it, it is of interest to examine the distribution of the
 6    levels of 1-hour PM2 5 that result from the method. Ideally, after rollback any area that had a
 7    non-zero required reduction should have a post-rollback design value for 1-hour PM2.5 mass that
 8    is exactly equal to the target design value. Also, there should be a progression of reductions in 1-
 9    hour PM2.5 medians and other percentile points on the distribution as progressively more
10    stringent scenarios are modeled.
11
12       Table A-5 shows the post-rollback 1-hour PM2.5 mass design values for the scenarios, with
13    percentile forms matched. Design values for area-scenario combinations for which the required
14    reductions were zero have been omitted, because the current conditions design values for these
15    combinations would not be expected to reflect the target design value.  It can be seen that the
16    design values progress as expected and are in the vicinity of the target design values, but are not
17    always exactly equal to the targets. EPA staff attributes this  to the fact that PRB concentrations
18    of 1-hour PM2 5 mass vary from hour to hour.  It is possible for the daily maximum PM2 5 mass
19    concentration on a certain day in 2005 with a percentile rank of, for example, 96th to have a
20    relatively small PRB portion and a large non-PRB portion compared to the daily maximum
21    concentration that ranks 95th.  When an equal reduction is made to the non-PRB portion of each
22    total concentration, the two values may switch rank positions, and so a new day and hour
23    becomes the 2005 contributor to the rolled back three-year design value.  Since this day and hour
24    was not used to determine the required percentage reduction, the resulting design value will not
25    exactly meet the target design value.  It would be possible to iterate with higher and  lower
26    percentage reductions until the rolled back design value exactly matched the target design value,
27    but EPA considered this degree of refinement to be unnecessary in order to meet the objectives
28    of the Policy Assessment Document,  given other uncertainties in the underlying data and in the
29    assumptions used to estimate PM light extinction values.
30
31    EPA staff also generated and examined box plots of daily maximum daylight 1-hour PM2 5 mass
32    concentrations as a check for conceptual or programming, and found them to match  expectations.
33    They are not included here, for conciseness.
     Draft - Do Not Quote or Cite                A-8                               March 2010

-------
1
2
3
4
5
     Table A-5. Post-rollback design values for daily maximum 1-hour PM2.s mass. Design
     values are shown only for combinations of study area and scenario for which the study area
     does not meet the scenario under current conditions, such that reductions were made
     during the rollback modeling.
Scenario
Level (ug/m3 )
Statistical Form
Study Area
Tacoma
Fresno
Los Angeles
Phoenix
Salt Lake City
Dallas
Houston
St. Louis
Birmingham
Atlanta
Detroit
Pittsburgh
Baltimore
Philadelphia
New York
aa
60
90th
bb
60
95th
cc
40
90th
dd
40
95th
ee
30
90th
ff
30
95th
gg
20
90th
hh
20
95th
ii
10
90th
ii
10
95th
Corresponding Design Value (ug/m3 ) (same percentile form as the scenario)


53







52


46


63
53





58

59


44


40
35





42

34
33

31
42

42
35

38


39
39
36
39
33
38
30
40

30
26

29

29
31
32
28
26
24
31
23
32

31
26

28

27
29
29
27
29
25
28
22
30
20
20
18

19
23
19
21
21
19
17
16
21
16
21
21
21
18
19
19
23
18
19
20
18
20
17
19
15
20
11
10
9
10
10
12
10
10
11
10
9
8
10
8
11
12
10
9
10
10
11
9
10
10
10
10
9
10
8
10
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
    6.0 PM Light Extinction Results for "Just Meeting" Alternative Secondary NAAQS
    Scenarios Based on 1-hour PM2.s Mass

        The rollback steps described in section 4.0 resulted in estimates of PM light extinction for
    each day and hour in each study area, for each of the 10 NAAQS scenarios based on 1-hour
    PM2.5 mass as the indicator.  Two summaries of these conditions are presented here.

        Figure A-2 presents a box plot of daily maximum daylight 1-hour PM light extinction for
    each NAAQS scenario based on 1-hour PM2 5 mass. These can be compared to Figure 3-8(a) of
    the UFVA representing pre-rollback daily maximum PM light extinction, and to the upper panel
    of the figures in Appendix F of the UFVA representing the daily maximum PM light extinction
    levels resulting from the 20 NAAQS scenarios examined in the UFVA (18 scenarios based on
    PM light extinction as the indicator, the current annual and 24-hour PM2 5 NAAQS, and a
    scenario with an annual NAAQS of 12 |ig/m3 and a 24-hour NAAQS  of 25 |ig/m3). It can be
    seen that the distribution of PM2 5 mass in a given study area shifts downward as the NAAQS
    scenarios progress from least to most stringent (as indicated by the required percentage
    reduction) and in most cases become more similar to other areas (once the progression of
    scenarios begins to require reductions in a given area). St.  Louis is an obvious exception, in that
    it retains many relatively high values even under the most  stringent NAAQS scenario. This is
    due to the fact that many hours and days in St. Louis have  a large estimated PMio-2.s
    concentration contribution to estimated PM light extinction, as visualized in the light extinction
    budgets presented in Figure 3-13 of the UFVA. The rollback approach does not change PMio-2.s
    concentrations.  Recall from section 3.3.2 and Appendix A of the UFVA that for St. Louis,
     Draft - Do Not Quote or Cite
                                             A-9
March 2010

-------
 1    2.5 concentrations were estimated based on a PMio instrument located very near a
 2    recycling/municipal works yard and a PM2 5 instrument in another less industrial location. As
 3    such, those estimated concentrations of PMio-2.s may not represent a large portion of the St Louis
 4    urban area.
 5
 6       PMio-2.5 was also a notable contributor to the estimated light extinction budgets in Los
 7    Angeles on a number of days, although to a lesser degree than in St. Louis. This is reflected in
 8    the box plots, which show a number of high values of PM light extinction regardless of the level
 9    of the hypothetical PM2.5 mass NAAQS. Recall from section 3.3.2 and Appendix A of the
10    UFVA that for Los Angeles, PMi0-2.5 concentrations were estimated based on a PMio instrument
11    located in Victorville, California which is a considerable distance from the site in Rubidoux,
12    California and is perhaps a dustier area than most of the Los Angeles airshed.
13
14       Table A-6 presents the percentage of days in 2005-2007 on which daily maximum 1-hour
15    PM light extinction exceeded each of the CPL, under each of the 10 secondary PM NAAQS
16    scenarios based on 1-hour PM2.s mass. These percentages are necessarily based on the days for
17    which data to estimate PM light extinction were available, but are best estimates of the
18    percentage of all days in the year given that the days with data were well distributed across the
19    year on either a one-in-three or one-in-six sampling schedule.  These percentages can be
20    compared to the same-basis percentages presented in Table 4-7 of the UFVA.
     Draft - Do Not Quote or Cite               A-10                              March 2010

-------
1
2
3
4
5
6
7
10
11
12
    Figure A-2. Distributions of maximum daily daylight 1-hour PM light extinction under
    "just meet" conditions for NAAQS scenarios based on 1-hour PM2.s mass (excluding hours
    >90% RH)

    (aa) NAAQS Scenario
    60 ng/m3
     90th percentile
                      PMRolfcackDailyMaxNAAOS60PcII90DVsFromdaily.max
              r "
                      |i4mii:HI
                      j__(.l_j-J.J___l...-|r.	j	1	.|......^...L.__J...^_....  .
                 -" ,* f   *' S
    (bb) NAAQS Scenario
    60 jig/m3
    95th percentile
                      PMRollbackDaityMaxNAAOS60Pcll95DVsFrom
-------
 1
 2
 3
 4
 5
 6
 7
10
1 1
12
     Figure A-2. Distributions of maximum daily daylight 1-hour PM light extinction under
     "just meet" conditions for NAAQS scenarios based on 1-hour PM2.s mass (excluding hours
     >90%RH) (continued)

     (cc) NAAQS Scenario
     40 ng/m3
      90th percentile
                           PMRollbackDaiVMaxNMQS40Pcil90DVsFroTOteily.max
             4-
           =g=3
                                                 .1.
                                                     1
-1-
     (dd) NAAQS Scenario
     40^g/m3
      95th percentile
                           PMRollbackDaityMaxNAAOS40Pell95DVsFrcmxJaily.max
        1-
                     I
                                       0
                                                  io              IS
                                               .  iiii-A
                                                -1	U	1

     Draft - Do Not Quote or Cite
                                              A-12
                     March 2010

-------
1
2
3
4
5
6
7
      Figure A-2.  Distributions of maximum daily daylight 1-hour PM light extinction under

      "just meet" conditions for NAAQS scenarios based on 1-hour PM2.s mass (excluding hours

      >90% RH) (continued)


      (ee) NAAQS Scenario
      30 ng/m3
      90th percentile

                              PMRollbackDaiVMaxNMQS30Pcil90DVsFromdaily.max
      (ff) NAAQS Scenario
10
11    95th percentile
                              PMRollbackDaityMaxNAAOS30Pell95DVsFrcmxJaily.max
                                           a

                                          1
                                                      •
                                                  A-T
12
      Draft - Do Not Quote or Cite
                                                  A-13
March 2010

-------
1
2
3
4
5
6
7
10
11
12
      Figure A-2. Distributions of maximum daily daylight 1-hour PM light extinction under
      "just meet" conditions for NAAQS scenarios based on 1-hour PM2.s mass (excluding hours
      >90% RH) (continued)

      (gg) NAAQS Scenario
         20 ng/m3
      90th percentile
                              PMRo)lbac*DailyMaxNMOS20Pcll90DVsFrom
-------
1
2
3
4
5
6
7
8
 9
10
11
12
13
14
      Figure A-2. Distributions of maximum daily daylight 1-hour PM light extinction under
      "just meet" conditions for NAAQS scenarios based on 1-hour PM2.s mass (excluding hours
      >90% RH) (continued)
      (ii) NAAQS Scenario
      Daily Max
      90th percentile
        §-1
     (jj) NAAQS Scenario
     Daily Max
     95th percentile
                           PMRollbackDailyMaxNAAQSlOPcll95DVsFromdaily.max
                      wote or C/te
                                                A-15
March 2010

-------
1
2
3
4
5
    Table A-6.  Percentage of days across three years (two in the case of Phoenix and Houston) with maximum 1-hour daylight PM
    light extinction above CPLs when "just meeting" NAAQS scenarios based on 1-hour PM2.s mass. Blue shading indicates no
    reduction required from current conditions.


Scenario
NAAQS
Level
(Mg/m3)
NAAQS
Percentile
Form
Area
Tacoma
Fresno
Los Angeles
Phoenix
Salt Lake
City
Dallas
Houston
St. Louis
Birmingham
Atlanta
Detroit
Pittsburgh
Baltimore
Philadelphia
New York
Average
Days with max hour above
64 Mm *
aa bb cc dd ee ff gg hh ii jj
60 60 40 40 30 30 20 20 10 10
90 95 90 95 90 95 90 95 90 95
Percentage of days
53 53 53 53 53 53 53 53 43 35
76 73 65 57 69 60 55 44 28 17
89 87 84 81 84 79 74 69 41 30
44 44 44 44 44 44 44 44 37 32

45 45 45 37 45 45 45 26 17 10
81 81 81 81 81 81 81 71 41 29
79 79 79 79 79 79 74 65 32 27
98 98 98 97 98 97 95 89 73 67
89 85 80 68 87 80 72 62 41 34
91 91 91 89 91 89 82 77 47 34
84 80 74 72 76 73 65 60 40 33
85 85 81 77 81 77 63 55 27 19
81 81 81 76 81 74 64 56 31 20
84 78 71 62 72 63 55 43 17 10
83 83 80 71 81 73 63 56 27 19
78 76 74 70 75 71 66 58 36 28
Days with max hour above
112 Mm -1
aa bb cc dd ee ff gg hh ii jj
60 60 40 40 30 30 20 20 10 10
90 95 90 95 90 95 90 95 90 95
Percentage of days
23 23 23 23 23 23 23 23 11 6
52 48 37 31 44 32 29 18 9 4
78 76 65 57 65 53 41 31 11 7
6 66 6666 666

17 17 17 15 17 17 17 11 85
41 41 41 41 41 41 41 32 8 5
44 44 44 44 44 44 35 28 6 3
78 78 78 74 78 75 64 55 34 29
65 56 51 36 58 51 40 30 15 12
75 75 75 68 74 66 51 35 3 3
67 57 51 43 53 48 34 21 96
57 57 51 45 52 44 29 22 3 0
51 51 51 45 51 44 31 23 4 3
60 54 33 29 37 31 16 10 3 3
60 60 56 39 56 40 32 22 6 3
52 49 45 40 47 41 33 25 9 6
Days with max hour above
191 Mm -1
aa bb cc dd ee ff gg hh ii jj
60 60 40 40 30 30 20 20 10 10
90 95 90 95 90 95 90 95 90 95
Percentage of days
4 44444441 1
30 27 17 11 23 12 10 5 1 0
52 46 30 24 30 19 11 6 3 3
1 11111111 1

8 88788852 1
10 10 10 10 10 10 10 7 0 0
11 11 11 11 11 11 9 6 1 0
40 40 40 36 40 38 27 20 13 13
34 26 21 13 30 20 15 11 4 3
31 31 31 21 31 19 5 3 0 0
43 28 13 7 14 9 6 41 1
26 26 18 14 21 13 6 2 0 0
23 23 23 18 23 16 8 2 1 1
26178585000 0
29 29 25 16 25 17 9 5 0 0
24 22 17 13 19 14 9 5 2 2
6
7
          Draft Do Not Quote or Cite
                                                           A-16
March 2010

-------
 1             5   REVIEW OF THE SECONDARY STANDARDS FOR OTHER
 2                                        WELFARE EFFECTS

 3           This chapter presents preliminary staff conclusions with regard to the current suite of
 4    secondary PM standards to protect against PM-related welfare effects other than visibility
 5    impairment.  Specifically, staff has assessed the relevant information related to effects of
 6    atmospheric PM on the environment, including effects on climate, ecological effects, and effects
 7    on materials.  Our assessment is framed by a series of key policy-relevant questions, which
 8    expand upon those presented in the Integrated Review Plan (IRP) (US EPA, 2008a, section 3.2).
 9    The answers to these questions will inform decisions on whether to retain or revise the current
10    suite of secondary PM standards.
11           In presenting preliminary staff conclusions with regard to the current secondary standards
12    relative to PM-related effects on climate, ecological effects, and materials, we note that the final
13    decision is largely a public welfare policy judgment.  A final decision must draw upon scientific
14    information and analyses about non-visibility PM-related effects and related impacts on public
15    welfare, as well as judgments about how to deal with the range of uncertainties that are inherent
16    in the scientific evidence and analyses. Our approach to informing these judgments is discussed
17    more fully below. This approach is consistent with the requirements of the NAAQS provisions
18    of the Act and with how EPA and the courts have historically interpreted the Act. These
19    provisions require the Administrator to establish secondary standards that, in the Administrator's
20    judgment, are requisite to protect public welfare from any known or anticipated adverse effects
21    associated with the presence of the pollutant in the ambient air.  In so doing, the Administrator
22    seeks to establish standards that are neither more nor less stringent than necessary for this
23    purpose.  The Act does not require that secondary standards be set at a zero-risk level, but rather
24    at a level that avoids unacceptable  public welfare impacts.
25           Information on the approaches used to set the secondary PM standards in past reviews as
26    well as our current approach for this review are presented in section 5.1.  A discussion of the
27    scope of the review as related to non-visibility welfare effects of PM is included in section 5.1.2.
28    This chapter considers each of the  non-visibility welfare effects separately. The discussion of
29    PM-associated effects on climate (section 5.2), ecological effects (section 5.3), and materials
30    (section 5.4) are each followed by  a consideration of key uncertainties and areas for future
31    research and data collection.
                                                 5-1
      Draft Do Not Quote or Cite                                             March 2010

-------
 1    5.1    APPROACH
 2          Background information on the approaches used to establish the PM secondary standards
 3    in 1997 and revisions to those standards in 2006 are summarized below.  This section also
 4    includes a discussion of the ongoing j oint review of ecological effects of oxides of nitrogen and
 5    sulfur (NOx/SOx secondary review) for clarity, since deposit!onal effects of PM components of
 6    NOx and SOx to ecosystems were historically considered as a component of the PM secondary
 7    review.  Lastly, there is a discussion of the current approach for evaluating the effects of PM on
 8    climate,  ecosystems, and materials using evidence-based considerations to inform our
 9    understanding of the key policy-relevant issues.

10    5.1.1  Approaches Used in Previous Reviews
11          5.1.1.1 Review Completed in  1997
12          In the 1997 review, as discussed in section 2.1.1.1, EPA determined that for the primary
13    standard the fine and coarse fractions of PMio should be considered separately and added a suite
14    of new primary standards, using PM2.s, as the indicator for fine particles, and retaining PMio as
15    the indicator for regulating thoracic coarse particles.  The EPA established two new PM2.5
16    standards: an annual standard of 15 |ig/m3, based on the 3-year average of annual arithmetic
17    mean PM2.5 concentrations from single or multiple community-oriented monitors; and a 24-hour
18    standard of 65 |ig/m3, based on the 3-year average of the 98th percentile of 24-hour PM25
19    concentrations at each population-oriented monitor within an area (62 FR 38652, July 18, 1997).
20          With respect to the secondary PM standards, EPA concluded in 1997, that the available
21    evidence on effects of PM on non-visibility welfare endpoints was not sufficient to warrant a
22    separate secondary standard. Therefore, the secondary standards were set equal to the primary
23    PM2.5 and PMio standards in the final rule to provide protection against effects on visibility as
24    well as materials damage and soiling effects related to fine and coarse particles (62 FR 38683).

25          5.1.1.2 Review Completed in  2006
26          In 2006, the Administrator concluded that there was insufficient information to consider a
27    distinct secondary standard based on PM-related impacts to ecosystems, materials damage and
28    soiling, and climatic and radiative processes (71  FR 61144, October 17, 2006).  Specifically,
29    there was a lack of evidence linking various non-visibility welfare effects to specific levels of
30    ambient PM. To provide a level of protection for welfare-related effects, the secondary
31    standards were set equal to the revised primary standards to directionally improve the level of
32    protection afforded vegetation, ecosystems and materials (71 FR 61210).
                                                5-2
     Draft Do Not Quote or Cite                                              March 2010

-------
 1          In the last review, the 2004 AQCD concluded that regardless of size fraction, particles
 2   containing nitrates and sulfates have the greatest potential for widespread environmental
 3   significance (US EPA, 2004, sections 4.2.2 and 4.2.3.1). Considerable supporting evidence was
 4   available that indicated a significant role of NOx,  SOx, and transformation products in
 5   acidification and nutrient enrichment of terrestrial and aquatic ecosystems (71 FR 61209). The
 6   recognition of these ecological effects, coupled with other considerations detailed below, led
 7   EPA to initiate a joint review of the NC>2 and SC>2 secondary NAAQS that will consider the
 8   gaseous and particulate species of NOx and SOx with respect to the ecosystem-related welfare
 9   effects that result from the deposition of these pollutants and transformation products.

10   5.1.2  Scope of Current NAAQS Reviews
11          Non-visibility welfare-based effects of oxides of nitrogen and sulfur are divided between
12   two NAAQS reviews; (1) the PM NAAQS review and, (2) the joint NOx/SOx secondary
13   NAAQS review.  The scope of each document and the components of N and S considered in
14   each review are detailed in this section and summarized in Table 5-1.
                                                5-3
     Draft Do Not Quote or Cite                                             March 2010

-------
1
2
3
Table 5-1. Scope of the current secondary PM NAAQS review and current NOx/SOx secondary review.

Welfare
Effect
Documents
ISA
REA
PA
Components
NOx/SOx Secondary
Review
Acidifying
deposition,
nutrient
enrichment
NOx/SOx
NOx/SOx
NOx/SOx
Deposited
particulate and
gaseous forms
of oxides of
nitrogen and
sulfur and
related N and S
containing
compounds.
Direct effects
of gas-phase
NOx/SOx on
vegetation
NOx/SOx
NOx/SOx
NOx/SOx
Gaseous forms
of oxides of
nitrogen and
sulfur and
related N and S
containing
compounds in
the ambient air.
PM Secondary
Review
Visibility
impairment
PM
PM (Urban
focused
visibility
assessment)
PM
All particles
10 microns or
smaller in the
ambient air.
Climate
Forcing effects
PM

PM
Climate-related
particles
(aerosols) in
the ambient air.
Ecological
effects
PM

PM
Deposited
components
ofPM,
including
metals and
organics but
not N and S
containing
compounds.
Materials
Damage
PMand
NOx/SOx
Annex E

PM
Particles and
gases
associated with
ambient NOx
and SOx
including NOy,
NH3 and NHx.
Soiling
PM

PM
Deposited
particles
    Draft Do Not Quote or Cite
                                                          5-4
                                                               March 2010

-------
 1          5.1.2.1  Scope of the Current Secondary PM NAAQS Review
 2          In reviewing the current suite of secondary PM standards to address visibility impairment
 3    (chapter 4), climate forcing effects (section 5.2), and other welfare-related effects (sections 5.3
 4    and 5.4), all PM-related effects that are not being covered in the NOx/SOx review are
 5    considered. With regard to the materials section (5.4), the discussion has been expanded to
 6    include particles and gases that are associated with the presence of ambient NOx and SOx, as
 7    well as NOy, NHa and NHx for completeness.  By excluding the effects associated with
 8    deposited particulate matter components of NOx and SOx and their transformation products
 9    which are addressed fully in the NOx/SOx secondary review, as outlined below, the  discussion
10    of ecological effects of PM has been narrowed to focus on effects associated with the deposition
11    of metals and, to a lesser extent, organics (section 5.3).

12          5.1.2.2 Scope of the Current NOx/SOx Secondary NAAQS Review
13          This is the first time since the NAAQS were established in 1971 that a joint review of the
14    secondary NAAQS  for NOx and SOx, has been conducted. This review is being conducted
15    because the atmospheric chemistry and environmental effects of NOx,  SOx, and their associated
16    transformation products are linked, and because the National Research Council (NRC) has
17    recommended that EPA consider multiple pollutants, as appropriate,  in forming the scientific
18    basis for the NAAQS.  The NOx/SOx secondary review focuses on the welfare effects  associated
19    with exposures from deposited parti culate and gaseous forms of oxides of nitrogen and sulfur
20    and related N and S containing compounds and transformation products on ecosystem  receptors.
21    An assessment of the complex ecological effects associated with N deposition requires
22    consideration of multiple forms of N. These include evaluation of data on inorganic reduced
23    forms of N (e.g., ammonia [NHa] and ammonium ion  [NH4+]), inorganic oxidized forms (e.g.,
24    NOx, nitric acid [HNO3], nitrous oxide [N2O], nitrate  [NO3-]), and organic N compounds (e.g.,
25    urea, amines, proteins, nucleic acids). In addition to  acidification and N-nutrient enrichment,
26    other welfare effects related to deposition of N-and S-containing compounds are discussed, such
27    as SOx interactions with mercury (Hg) methylation. In addition, the NOx/SOx secondary review
28    includes evidence related to direct ecological effects of gas-phase NOx and SOx since  the direct
29    effects of gas-phase SOx on vegetation formed a primary basis for the initial establishment of the
30    secondary NAAQS  for SO2.
31          Effects of acidifying deposition associated with paniculate N and S are covered in the
32    recent Integrated Science Assessment for Oxides of Nitrogen and Sulfur-Ecological Criteria
33    (Final Report (US EPA, 2008c). The Risk and Exposure Assessment for Review of the
34    Secondary National Ambient Air Quality Standards for Oxides of Nitrogen and Oxides of Sulfur
                                                5-5
      Draft Do Not Quote or Cite                                             March 2010

-------
 1    (Final)(NOx/SOx REA) (US EPA, 2009h) considers four main targeted ecosystem effects
 2    considered in the review of secondary effects of NOx and SOx: (1) aquatic acidification due to N
 3    and S, (2) terrestrial acidification due to N and S, (3) aquatic nutrient enrichment, including
 4    eutrophi cation and (4) terrestrial nutrient enrichment. In the draft Policy Assessment for Review
 5    of the Secondary National Ambient Air  Quality Standards for Oxides of Nitrogen and Oxides of
 6    Sulfur (US EPA 2010c) ecologically-based indicators that link atmospheric concentrations to
 7    deposition are being considered.

 8    5.1.3  Current Approach
 9          The remainder of this chapter summarizes and highlights key aspects of the policy
10    relevant information from the ISA to help inform the Administrator's judgments regarding the
11    adequacy of the current suite of secondary PM NAAQS in relation to climate processes,
12    ecological effects, and materials damage.  The ISA uses a five-level hierarchy that classifies the
13    weight of evidence for causation, not just association, into a qualitative statement about the
14    overall weight of evidence and causality (US EPA, 2009a, section 1.5.5, Table 1-3):  causal
15    relationship;  likely to  be a causal relationship; suggestive of a causal relationship; inadequate to
16    infer a causal relationship; not likely to be a causal relationship (see US EPA, 2009a, Table 1-3).
17          Staff is evaluating evidence-based considerations primarily by assessing the evidence of
18    associations identified in the ISA. All relationships between PM and climate, ecological effects,
19    and materials damage effects identified in the ISA are considered to be either "likely causal" or
20    "causal".  The staffs approach in this review of non-visibility welfare effects of PM is to
21    consider information regarding particulate matter effects on climate, ecological endpoints and
22    materials.  This includes new literature  available since the last review as well as existing,
23    relevant information as presented in the ISA (US EPA 2009a).

24    5.2    CLIMATE
25    5.2.1  Scope
26           Information and conclusions about what is currently known about the role of PM in
27    climate is summarized in Chapter 9 of the PM ISA (US EPA, 2009a). The ISA concludes; "that
28    a causal relationship exists between PM and effects on climate, including both direct effects on
29    radiative forcing and indirect effects that involve cloud feedbacks that influence precipitation
30    formation and cloud lifetimes" (US EPA,  2009a, section 9.3.10).  Material from the climate
31    section of the ISA is principally drawn  from the U.S. Climate Change Science Program
32    Synthesis and Assessment Product 2.3,  Atmospheric Aerosol Properties and Climate Impacts., by
33    Chin et al., (CCSP 2009) and Chapter 2, Changes in Atmospheric Constituents and in Radiative
34    Forcing,(ForstQr et al., 2007) in the comprehensive Working Group I report in the Fourth
                                                5-6
      Draft Do Not Quote or Cite                                             March 2010

-------
 1   Assessment Report (AR4) from the Intergovernmental Panel on Climate Change (IPCC), Climate
 2   Change 2007: The Physical Science Basis.  Sections 9.3.7 (Fire as a Special Source of PM
 3   Welfare Effects), 9.3.9 (Other Special Sources and Effects), 9.3.9.1 (Glaciers and Snowpack)
 4   and 9.3.9.3 (Effects on Local and Regional Climate) of the ISA were written by NCEA staff.
 5   This section of the PA summarizes and synthesizes the policy-relevant science in the ISA for the
 6   purpose of helping to inform consideration of climate aspects in the review of the secondary PM
 7   NAAQS.
 8          Atmospheric PM (referred to as aerosols1 in the remainder of this section to be consistent
 9   with the ISA) affects multiple aspects of climate. These include absorbing and scattering of
10   incoming solar radiation, alterations in terrestrial radiation, effects on the hydrological cycle, and
11   changes in cloud properties (US EPA, 2009a,  section 9.3.1). Major aerosol components that
12   contribute to climate processes include black carbon (BC), organic carbon (OC), sulfates, nitrates
13   and mineral dusts. There is a considerable ongoing research effort focused on understanding
14   aerosol contributions to changes in global mean temperature and precipitation patterns.  The
15   Climate Change Research Initiative identified research on atmospheric concentrations and effects
16   of aerosols as a high research priority (National Research Council, 2001) and the IPCC 2007
17   Summary for Policymakers states that anthropogenic contributions to aerosols remain the
18   dominant uncertainty in radiative forcing (IPCC 2007). The current state of the science of
19   climate alterations attributed to PM is in flux as  a result of continually updated information.

20   5.2.2  Adequacy of the Current Standard
21          In considering the adequacy of the suite of secondary standards, staff addresses the
22   following overarching question:
23
24
25
Does currently available scientific information, as reflected in the ISA, support or call into
question the adequacy of the protection for climate effects afforded by the current suite of
                               secondary PM standards?
26          To inform the answer to this overarching question, staff has posed specific questions to
27    aid in assessing the available scientific evidence as related to climate effects attributed to
28    aerosols.  In considering the currently available scientific and technical information, we included
29    both the information available from the last review and information that is newly available since
30    the last review synthesized in Chapter 9 of the ISA (US EPA, 2009).
31    •   What new techniques are available to improve our understanding of climate effects of
32       aerosols?
       In the sections of the ISA included from IPCC AR4 and CCSP SAP2.3, 'aerosols' is more frequently used than
      "PM" and that word is retained.

                                                5-7
      Draft Do Not Quote or Cite                                             March 2010

-------
 1          Global climate change has increasingly been the focus of intense international research
 2    endeavors. Major efforts are underway to understand the complexities inherent in atmospheric
 3    aerosol interactions and to decrease uncertainties associated with climate estimations. Two recent
 4    reports, the US CCSP Product 2.3 and sections of the IPCC AR4 were combined to form the
 5    climate discussion in the ISA (CCSP 2009; Forster et al., 2007).  A review of the most recently
 6    available techniques for assessing climate-aerosol relationships is presented in the ISA. Aerosol
 7    measurement capabilities reviewed in the ISA include a discussion of the increasingly
 8    sophisticated instrumentation and techniques available for quantifying aerosols, the enhanced
 9    sensing capabilities of satellites, development of remote sensing networks and synergy of
10    measurements with model simulations (US EPA 2009a, section 9.3.2).  Advances in measured
11    aerosol properties as related to modeling as well as outstanding issues remaining in these
12    measurement-based studies are  elaborated in the ISA (US EPA 2009a sections 9.3.3 and 9.3.4).
13    Section 9.3.6 of the ISA, "Global Aerosol Modeling" considers the capabilities of climate
14    modeling that have developed over the last decade and limitations of the techniques currently in
15    use (US EPA 2009a).
16    •   To what extent does newly available evidence improve our understanding of the nature
17       and magnitude of climate  responses to PM (aerosols)?
18          Aerosols have direct and indirect effects  on climate processes. The direct effects of
19    aerosols on climate result mainly from particles  scattering light away from earth into space,
20    directly altering the radiative balance of the Earth-atmosphere system. This reflection of solar
21    radiation back to  space decreases the transmission of visible radiation to the surface of the earth
22    and results in a decrease in the heating rate of the surface and the lower atmosphere.  At the same
23    time, absorption of either incoming solar radiation or outgoing terrestrial radiation by particles,
24    primarily BC, results in an increased heating rate in the lower atmosphere.  Global estimates of
25    aerosol direct radiative forcing (RF) were recently summarized using a combined model-based
26    estimate (Forster et al., 2007). The overall, model-derived aerosol direct RF was estimated in the
27    IPCC AR4 as -0.5 (-0.9 to -0.1) watts per square meter (W/m2), with an overall level of scientific
28    understanding of this effect as "medium low" (Forster et al., 2007), indicating a net cooling
29    effect in contrast  to greenhouse gases (GHGs) which have a warming effect.
30          The contribution of individual aerosol components to total aerosol direct radiative forcing
31    is more uncertain than the global average (US EPA, 2009a, section 9.3.6.6).  The direct effect of
32    radiative scattering by atmospheric particles exerts an overall net cooling of the atmosphere,
33    while particle absorption of solar radiation leads to warming. For example, the presence of OC
34    and sulfates decrease warming from sunlight by  scattering shortwave radiation back into space.
35    Such a perturbation of incoming radiation by anthropogenic aerosols is designated as aerosol
36    climate forcing, which is distinguished from the  aerosol radiative effect of the total aerosol
                                                5-8
      Draft Do Not Quote or Cite                                              March 2010

-------
 1    (natural plus anthropogenic). The aerosol climate forcing and radiative effect are characterized
 2    by large spatial and temporal heterogeneities due to the wide variety of aerosol sources, the
 3    spatial non-uniformity and intermittency of these sources, the short atmospheric lifetime of
 4    aerosols (relative to that of the greenhouse gases), and processing (chemical and microphysical)
 5    that occurs in the atmosphere. For example, OC can be warming (positive forcer) when
 6    deposited on or suspended over a highly reflective surface such as snow or ice but, on a global
 7    average, is a negative forcer in the atmosphere.
 8           More information has also become available on indirect effects of aerosols.  Particles in
 9    the atmosphere indirectly affect both cloud albedo (reflectivity) and cloud lifetime by modifying
10    the cloud amount, and microphysical and radiative properties (US EPA, 2009a, section 9.3.6.4).
11    The RF due to these indirect effects (cloud albedo effect) of aerosols is estimated in the IPCC
12    AR4 to be -0.7(-1.8 to -0.3) W/m2 with the level of scientific understanding of this effect as
13    "low" (Forster et al., 2007).   Aerosols act as cloud condensation nuclei (CCN) for cloud
14    formation. Increased particulates in the atmosphere available as CCN with no change in
15    moisture content of the clouds have resulted in an increase in the number and decrease in the
16    size of cloud droplets in certain clouds that can increase the albedo of the clouds (the Twomey
17    effect).  Smaller particles slow the onset of precipitation and prolong cloud lifetime.  This effect,
18    coupled with changes in cloud albedo, increase the reflection of solar radiation back into space.
19    The altitude of clouds also effects cloud radiative forcing. Low clouds reflect incoming sunlight
20    back to space but do not effectively trap outgoing radiation, thus, cooling the planet, while higher
21    elevation clouds reflect some sunlight but more effectively can trap outgoing radiation and act to
22    warm the planet (US EPA, 2009a, section 9.3.3.5).
23           The total negative RF due to direct and indirect effects of aerosols computed from the top
24    of the atmosphere, on a global average, is estimated at -1.3 (-2.2 to -0.5) W/m2 in contrast to the
25    positive RF of +2.9 (+3.2 to +2.6) W/m2 for anthropogenic GHGs (IPCC 2007, pg. 200).
26           The understanding of the magnitude of aerosol effects on climate has increased
27    substantially in the last decade. Data on the atmospheric transport and deposition of aerosols
28    indicate a significant role for PM components in multiple aspects of climate.  Aerosols can
29    impact  glaciers,  snowpack, regional water supplies, precipitation and climate patterns (US EPA
30    2009a section 9.3.9).  Aerosols deposited on ice or snow can lead to melting and subsequent
31    decrease of surface albedo (US EPA 2009a, section 9.3.9.2).  Aerosols are potentially important
32    agents of climate warming in the Arctic and other locations (US EPA, 2009a, section 9.3.9).
33    Incidental fires and biomass burning are being recognized as having a significant impact on
34    PM2.s concentrations and climate forcing. Intermittent fires can occur at large enough scales to
35    affect hemispheric aerosol concentrations (US EPA 2009a, section 9.3.7).

                                                5-9
      Draft Do Not Quote or Cite                                              March 2010

-------
 1          A series of studies available since the last review examine the role of aerosols on local
 2    and regional scale climate processes (US EPA, 2009a, section 9.3.9.3). Studies on the South
 3    Coast Air Basin (SCAB) in California indicate aerosols may reduce near-surface wind speeds,
 4    which, in turn reduce evaporation rates and increase cloud lifetimes. The  overall impact can be a
 5    reduction in local precipitation (Jacobson and Kaufmann, 2006). Conditions in the SCAB impact
 6    ecologically sensitive areas including the Sierra Nevadas. Precipitation suppression due to
 7    aerosols in California (Givati and Rosenfield, 2004) and other similar studies in Utah and
 8    Colorado found that orographic precipitation decreased by 15-30% downwind of pollution
 9    sources. Evidence of regional-scale impacts of aerosols on meteorological conditions in other
10    regions of the U.S. are lacking.
11    •   To what extent does the currently available information provide evidence of association
12       between specific PM constituents (i.e. BC, OC, sulfates) and climate-related effects?
13          Advances in the understanding of aerosol components and how they contribute to climate
14    change have enabled refined global forcing estimates of individual PM constituents. The global
15    mean radiative effect from individual components of aerosols was estimated for the first time in
16    the IPCC AR4 where they were reported to be (all in W/m2 units): -0.4 (+0.2) for sulfate, -0.05
17    (+0.05) for fossil  fuel-derived OC, +0.2 (+0.15) for fossil fuel derived BC, +0.03 (+0.12) for
18    biomass burning,  -0.1 (±0.1) for nitrates, and -0.1  (+0.2) for mineral dust (US EPA, 2009a,
19    section 9.3.10). Sulfate  and fossil fuel-derived OC cause negative forcing whereas BC causes
20    positive forcing because of its highly absorbing nature (US EPA, 2009a, 9.3.6.3). Although BC
21    comprises only a  small fraction of anthropogenic aerosol mass load and aerosol optical depth
22    (AOD), its forcing efficiency (with respect to either AOD or mass) is an order of magnitude
23    stronger than sulfate and particulate organic matter (POM), so its positive shortwave forcing
24    largely offsets the negative  direct forcing from sulfate and POM (IPCC, 2007; US EPA 2009a,
25    9.3.6.3). Global loadings for nitrates and anthropogenic dust remain very  difficult to estimate,
26    making the radiative forcing estimates for these constituents particularly uncertain (US EPA,
27    2009a, section 9.3.7).
28          Improved  estimates  of anthropogenic emissions of some aerosols, especially BC and OC,
29    have promoted the development of improved global emissions inventories and  source-specific
30    emissions factors  useful in climate modeling (Bond et al.  2004). Recent data suggests that BC is
31    one of the largest individual warming agents after carbon dioxide (CO2) and perhaps methane
32    (CH4) (Jacobson 2000; Sato et al., 2003; Bond and Sun 2005).  There are several studies
33    modeling BC effects on  climate and/or considering emission reduction measures on
34    anthropogenic warming  detailed in section 9.3.9 of the ISA. Fires release large amounts of BC,
35    CO2, CH4 and OC (US EPA, 2009a, section 9.3.7).

                                               5-10
      Draft Do Not Quote or Cite                                            March 2010

-------
 1    5.2.3   Preliminary Staff Conclusions
 2         •   Aerosols alter climate processes directly through radiative forcing and by indirect
 3             effects on cloud brightness, changes in precipitation and possible changes in cloud
 4             lifetimes.
 5         •   Individual components of aerosols differ in their reflective properties, and direction of
 6             climate forcing. Overall, aerosols have a net climate cooling effect.
 7         •   Most climate model simulations are based on global scale scenarios. These models
 8             may fail to consider the local variations in climate forcing due to emissions sources and
 9             local meteorological patterns.
10         •   Aerosols that are warming are co-emitted with aerosols that are cooling.  The relative
11             mix of these components will vary in areas across the U.S. and over time.
12           Collectively taking into consideration the responses to specific questions regarding the
13    adequacy of the current secondary PM standards for climate effects, we revisit the overarching
14    question: "does available scientific information,  as reflected in the ISA, support or call into
15    question the adequacy of the protection for climate effects afforded by the current suite of
16    secondary PM standards?" As an initial matter, we considered the appropriateness of the current
17    secondary standard defined in terms of PM2 5 and PMi0 indicators, for providing protection
18    against potential climate effects of aerosols. Newly available scientific information on climate-
19    aerosol relationships has improved our understanding of direct and indirect effects of aerosols
20    and aerosol properties.  The major aerosol components that contribute to climate processes
21    include BC, OC, sulfate, nitrate and mineral dusts.  These components vary in their reflectivity,
22    forcing efficiencies and even in the direction of climate forcing. The current standards that are
23    defined in terms of aggregate size mass cannot be expected to appropriately target controls on
24    components of fine and coarse particles that are related to climate forcing  effects. Thus, the
25    current mass-based PM2 5 and PMi0 secondary standards are not an appropriate or effective
26    means of focusing protection against PM-associated climate effects due to these differences in
27    components.
28           Overall, there is a net climate cooling associated with aerosols in the global atmosphere
29    (US EPA, 2009a,  section 9.2.10). Staff recognizes that some individual aerosol components,
30    such as BC, are positive climate forcers, whereas others, such as OC and sulfates, are negative
31    climate forcers. However, aerosols that are warming are co-emitted with aerosols that are
32    cooling. The relative mix of components will vary in areas across the U.S. and over time. Due to
33    the spatial and temporal heterogeneity  of PM components that contribute to climate forcing,
34    uncertainties in the measurement of aerosol components, inadequate consideration of aerosol
35    impacts in climate modeling, insufficient data on local and regional microclimate variations and
36    heterogeneity of cloud formations, it is not currently feasible to conduct a quantitative analysis
37    for the purpose of informing revisions  of the current NAAQS PM standard based on climate.
                                                5-11
      Draft Do Not Quote or Cite                                             March 2010

-------
 1    Based on these considerations, we reach the preliminary conclusion that there is insufficient
 2    information at this time to base a national ambient standard on climate impacts associated with
 3    current ambient concentrations of PM or its constituents2.

 4    5.2.4   Key Uncertainties and Areas for Future Research and Data Collection
 5           Although considerable progress is being made in estimating aerosol contributions to
 6    climate fluctuations, significant uncertainties remain that preclude consideration of climate
 7    effects as a basis for establishing a separate NAAQS secondary standard. A major impediment
 8    at this time to establishing a secondary standard for PM based on climate is the lack of accurate
 9    measurement of aerosol contributions, specifically quantification of aerosol absorption and
10    inability to separate the anthropogenic component from total aerosol forcing. Section 9.3.4 of
11    the ISA details the current limitations in aerosol measurement. Most measurement studies focus
12    on the sum of natural and anthropogenic contributions under clear sky conditions, however, this
13    scenario is simplistic when effects of cloud cover and differing reflective properties of land and
14    ocean are considered. Satellite measurements do not currently have the capability to distinguish
15    anthropogenic from natural aerosols. Due to a lack of data on the vertical distribution of aerosols,
16    above-cloud aerosols and profiles of atmospheric radiative heating are poorly understood (US
17    EPA, 2009a, section 9.3.4).
18           Another uncertainty in considering climate effects of PM in the NAAQS review is the
19    spatial and temporal heterogeneity of aerosols. In regions having high concentrations of
20    anthropogenic aerosols, aerosol forcing is greater than the global average, and can exceed
21    warming by GHGs, locally reversing the sign of the forcing (US EPA, 2009a, section 9.3.1).
22    Emissions of carbonaceous aerosols from intermittent fires and volcanic activity can further
23    complicate regional climate forcing estimates (US EPA, 2009a, sections 9.3.7 and 9.3.8).
24    Individual components of aerosols may either be positive or negative climate forcers.  Airborne
25    PM components may be directly emitted or undergo a variety  of physical and chemical
26    interactions and transformations.  These result in changes in particle size, structure and
27    composition which alter aerosol reflective properties. Aerosols can grow in size in the
28    atmosphere because ambient water vapor condenses on individual particles,  a phenomenon
29    known as hygroscopic growth (US EPA, 2009a, section 9.3.6.2). Atmospheric lifetimes of
30    individual aerosol components vary greatly confounding tracking source receptor relationships.
31           Improved representation of aerosols in climate models is essential to more accurately
32    predict the role of PM in climate forcing (US EPA, 2009a, section 9.3.6.7). The influence of
      2 Given the reasons discussed above, this conclusion would apply for both the secondary (welfare based) and the
      primary (health based) standards.

                                                5-12
      Draft Do Not Quote or Cite                                             March 2010

-------
 1    aerosols on climate is not yet adequately taken into account in computer predictions although
 2    considerable progress in being made in this area. For example, PM components underrepresented
 3    or missing from many models include nitrate aerosols and anthropogenic secondary aerosols (US
 4    EPA, 2009a, section 9.3.6.7). The modeling of aerosol indirect effects and absorption are
 5    difficult due to the high level of uncertainty associated with these climate factors.
 6           The interaction of PM with clouds remains a large source of uncertainty in climate
 7    estimates.  The interactions of aerosols with clouds and linkages between clouds and the overall
 8    climate system are complex and limit the feasibility of conducting quantitative analysis for the
 9    purpose of establishing a secondary PM standard based on welfare effects on climate processes.

10    5.3     ECOLOGICAL EFFECTS
11    5.3.1   Scope
12           Information on what is currently known about ecological effects of PM is summarized in
13    Chapter 9 of the ISA (US EPA 2009a). Four main categories of ecological  effects are identified
14    in the ISA: direct effects, effects of PM-altered radiative flux, indirect effects of trace metals and
15    indirect effects of organics. Exposure to PM for direct effects occur via deposition (e.g. wet, dry
16    or occult) to vegetation surfaces, while indirect effects occur via deposition to ecosystem soils or
17    surface waters where the deposited constituents of PM then interacts with biological organisms.
18    Both fine and coarse-mode particles may affect plants and other organisms; however, PM size
19    classes do not necessarily relate to ecological effects (U.S. EPA, 1996). More often the chemical
20    constituents drive the ecosystem response to PM (Grantz et al., 2003).  The trace metal
21    constituents of PM considered in the ecological effects section of the ISA are cadmium (Cd),
22    copper (Cu), chromium (Cr), mercury (Hg), nickel (Ni) and zinc (Zn). Ecological effects  of lead
23    (Pb) in particulate form are covered in the Air Quality Criteria Document for Lead (US EPA,
24    2006).  The organics included in the ecological effects section of the ISA are persistent organic
25    pollutants (POPs), polyaromatic hydrocarbons (PAHs) and polybromiated diphenyl ethers
26    (PBDEs).
27           Ecological effects of PM include direct effects to metabolic processes of plant foliage;
28    contribution to total metal loading resulting in alteration of soil biogeochemistry and
29    microbiology, plant and animal growth and reproduction; and contribution to total organics
30    loading resulting in bioaccumulation and biomagnification across trophic levels. It is important
31    to emphasize that the metal and organic constituents of PM contribute to total metal and organic
32    loads in ecosystems.
33           The ISA states that overall, ecological evidence is sufficient to conclude that a causal
34    relationship is likely to exist between deposition of PM and a variety of effects on individual

                                                5-13
      Draft Do Not Quote or Cite                                             March 2010

-------
 1    organisms and ecosystems based on information from the previous review and limited new
 2    findings in this review (US EPA 2009a, sections 2.5.3 and 9.4.7).  However the ISA also finds,
 3    in many cases, it is difficult to characterize the nature and magnitude of effects and to quantify
 4    relationships between ambient concentrations of PM and ecosystem response due to significant
 5    data gaps and uncertainties as well as considerable variability that exists in the components of
 6    PM and their various ecological effects.
 7           Ecological effects of PM must then be evaluated to determine if they are known or
 8    anticipated to have an adverse impact on public welfare. Characterizing a known or anticipated
 9    adverse effect to public welfare is an important component of developing any secondary
10    NAAQS. The most recent secondary NAAQS reviews have assessed changes in ecosystem
11    structure or processes using a weight-of-evidence approach that uses both quantitative and
12    qualitative data. For example, the 2008 ozone (63) final rule and 2010 63 proposal conclude that
13    a determination of what constitutes an "adverse" welfare effect in the context of secondary
14    NAAQS review can appropriately occur by considering effects at higher ecological levels
15    (populations, communities, ecosystems) as supported by recent literature. In the 2008
16    rulemaking and current ozone proposal, the interpretation of what constitutes an adverse effect
17    on vegetation can vary depending on the location and intended use of the plant. The degree to
18    which Os-related effects are considered adverse depends on the intended use of the vegetation
19    and its significance to public welfare (73 FR 16496). Therefore, effects (e.g. biomass loss, foliar
20    injury, impairment of intended use) may be judged to have a different degree of impact on public
21    welfare depending, for example, on whether that effect occurs in a Class I area, a city park,
22    commercial cropland or private land.
23           A paradigm useful in evaluating ecological adversity is the concept of ecosystem
24    services.  Ecosystem services identify the varied and numerous ways that ecosystems are
25    important to  human welfare. Ecosystems provide many goods and services that are of vital
26    importance for the functioning of the biosphere and provide the basis for the delivery of tangible
27    benefits to human  society.  An EPA initiative to consider how ecosystem structure and function
28    can be interpreted through an ecosystem services approach has resulted in the inclusion of
29    ecosystem services in the NOx/SOx REA (US EPA, 2009h). The Millennium Ecosystem
30    Assessment (MEA) defines these to include supporting, provisioning, regulating and cultural
31    services (Hassan et al., 2005):
32         •   Supporting services are necessary for the production of all  other ecosystem services.
33             Some examples include biomass production, production of atmospheric C>2, soil
34            formation and retention, nutrient cycling, water cycling, and provisioning of habitat.
35            Biodiversity is a supporting service that is increasingly recognized to sustain many of
36            the goods and services that humans enjoy from ecosystems. These provide a basis for
37            three higher-level categories of services.
                                               5-14
      Draft Do Not Quote or Cite                                            March 2010

-------
 1         •   Provisioning services, such as products (Gitay et al., 2001) i.e., food (including game,
 2             roots, seeds, nuts, and other fruit, spices, fodder), fiber (including wood, textiles), and
 3             medicinal and cosmetic products (including aromatic plants, pigments).
 4         •   Regulating  services that are of paramount importance for human society such as (a) C
 5             sequestration, (b) climate and water regulation, (c) protection from natural hazards
 6             such as floods, avalanches, or rock-fall,  (d) water and air purification, and (e) disease
 7             and pest regulation.
 8         •   Cultural services that satisfy human spiritual and aesthetic appreciation of ecosystems
 9             and their components.
10           An important  consideration in evaluating biologically adverse effects of PM and linkages
11    to ecosystem services is that  many of the MEA categories overlap and any one pollutant may
12    impact multiple services.  For example, deposited PM may alter the composition of soil-
13    associated microbial communities, which may affect supporting services such as nutrient
14    cycling. Changes in available soil nutrients could result in alterations to provisioning services
15    such as timber yield and regulating services such as climate regulation.  If enough information is
16    available, these alterations can be quantified based upon economic approaches for estimating the
17    value of ecosystem services.  Valuation may be important from a policy perspective because it
18    can be used to compare the benefits of altering versus maintaining an ecosystem.  Knowledge
19    about the relationships linking ambient concentrations and ecosystem services can be used to
20    inform a policy judgment on  a known or anticipated adverse public welfare effect.
21           This review seeks to build upon and focus this body of science using the concept of
22    ecosystem services to qualitatively evaluate linkages between biologically adverse effects and
23    particulate deposition. This approach is similar to that taken in the NOx/SOx REA in which the
24    relationship between  air quality indicators, deposition of N and S, ecologically relevant
25    indicators and effects on sensitive receptors are linked to changes in ecosystem structure and
26    services (US EPA, 2009h). This approach considers  the benefits received from the resources and
27    processes that are supplied by ecosystems.  Ecosystem components (e.g. plants, soils, water,
28    wildlife) are impacted by  PM air pollution, which may alter the services provided by the
29    ecosystems in question. The goals of this policy assessment are to (1) identify ecological effects
30    associated with PM deposition that can be linked to ecosystem services  and (2) qualitatively
31    evaluate ecological endpoints when possible. Keeping these goals and guidelines in mind,
32    limited new data on PM effects on plants, soil and nutrient cycling, wildlife and water are
33    evaluated in  the context of ecosystem services to qualitatively evaluate linkages between
34    biologically adverse effects and particulate deposition for the purpose of evaluating the adequacy
35    of the current standard.
                                                5-15
      Draft Do Not Quote or Cite                                              March 2010

-------
 1    5.3.2  Adequacy of the Current Standard
 2          In considering the adequacy of the suite of secondary standards, staff addresses the
 3    following overarching question:
 4
 5
 6
 Does available scientific information, as reflected in the ISA support or call into question
the adequacy of the protection afforded by the current suite of secondary PM standards for
 vegetation and ecosystems from the effects of deposited particulate metals and organics?
 7          To inform the answer to this overarching question, staff has posed specific questions to
 8    aid in assessing the available scientific evidence as related to ecosystem effects attributed to PM
 9    deposition as presented in the ISA (US EPA, 2009a).
10    •   To what extent has key scientific evidence become available to improve our
11       understanding of the nature and magnitude of ecosystem responses, the variability
12       associated with these responses, and the impact of PM on ecosystem services?
13          Key scientific evidence regarding PM effects on plants, soil and nutrient cycling, wildlife
14    and water available since the last review is summarized below to evaluate how this information
15    has improved  our understanding of ecosystem responses to PM.
16    Plants
17          As primary producers, plants play a pivotal role in energy flow through ecosystems.
18    Ecosystem services derived from plants include all of the categories (supporting, provisioning,
19    regulating, cultural) identified in the MEA (Hassan et al., 2005). Vegetation supports other
20    ecosystem processes by cycling nutrients through food webs and serving as a source of organic
21    material for soil formation and enrichment.  Trees and plants provide food, wood, fiber, and fuel
22    for human consumption. Flora help to regulate climate by sequestering CCh,  control  flooding by
23    stabilizing soils and cycling water via uptake and evapotranspiration.  Plants are significant in
24    aesthetic, spiritual and recreational  aspects of human interactions.
25          Particulate matter can adversely impact plants and ecosystem services provided by plants
26    by deposition  to vegetative surfaces (US EPA, 2009a, section 9.4.3). Particulates deposited on
27    the surfaces of leaves and needles can block light, altering the radiation received by the plant.
28    PM deposition can obstruct stomata limiting gas exchange, damage leaf cuticles and increase
29    plant temperatures.  This level of PM accumulation is typically observed near sources of heavy
30    deposition such as smelters and mining operations (US EPA, 2009a, section 9.4.3). Plants
31    growing on roadsides exhibit impact damage from near-road PM deposition, having higher levels
32    of organics and heavy metals, and accumulate salt from road de-icing during winter months (US
33    EPA, section 2009a, sections 9.4.3.1 and 9.4.5.7).
34          In addition to damage to plant surfaces, deposited PM can be taken up by plants from soil
35    or foliage. The ability of vegetation to take up heavy metals and organics is dependent upon the
36    amount, solubility and chemical composition of the deposited PM.  Uptake of PM by plants from
                                                5-16
      Draft Do Not  Quote or Cite                                              March 2010

-------
 1    soils and vegetative surfaces can disrupt photosynthesis, alter pigments and mineral content,
 2    reduce plant vigor, decrease frost hardiness and impair root development.  The ISA indicates that
 3    there are little or no effects on foliar processes at ambient levels of PM (sections 9.4.3 and 9.4.7)
 4    however, damage due to atmospheric pollution can occur near point-sources or under conditions
 5    where plants are subjected to multiple stressors.
 6          Though all heavy metals can be directly toxic at sufficiently high concentrations, only
 7    Cu, Ni, and Zn have been documented as being frequently toxic to plants (U.S. EPA, 2004),
 8    while toxicity due to Cd,  Co, and Pb has been observed less frequently (Smith, 1990; US EPA
 9    2009a, section 9.4.5.3). In general, plant growth is negatively correlated with trace metal and
10    heavy metal concentration in soils and plant tissue (Audet and Charest, 2007). Trace metals,
11    particularly heavy metals, can influence forest growth. Growth suppression of foliar microflora
12    has been shown to result from Fe, Al, and Zn. These three metals can also  inhibit fungal spore
13    formation, as can Cd, Cr, Mg, and Ni (see Smith,  1990). Metals cause stress and decreased
14    photosynthesis (Kucera et al., 2008) and disrupt numerous enzymes and metabolic pathways
15    (Strydom et al., 2006). Excessive concentrations of metals result in phytotoxicity through: (i)
16    changes in the permeability of the cell membrane; (ii) reactions of sulfydryl (-SH) groups with
17    cations; (iii) affinity for reacting with phosphate groups and active groups  of ADP or ATP; and
18    (iv) replacement of essential ions (Patra et al., 2004).
19          New information  since the last review provides additional evidence of plant uptake of
20    organics (US EPA, 2009a, section 9.4.6).  An area of active study is the impact of PAHs on
21    provisioning ecosystem services due to the potential for human and other animal exposure via
22    food consumption (US EPA, 2009a, section 9.4.6  page 9-190). The uptake of PAHs depends on
23    the plant species, site of deposition, physical and chemical properties of the organic compound
24    and prevailing environmental conditions.  It has been established that most bioaccumulation of
25    PAHs by plants occurs via leaf uptake, and to a lesser extent, through roots.  Differences
26    between species in uptake of PAHs confound attempts to quantify impacts to ecosystem
27    provisioning services. For example, zucchini (Cucurbitapepo) accumulated significantly more
28    PAHs than related plant species (Parrish et al., 2006).
29          Plants as ecosystem regulators can serve as passive monitors of pollution (US EPA,
30    2009a, section 9.4.2.3). Lichens and mosses are sensitive to pollutants associated with PM and
31    have been used with limited success to show spatial and temporal patterns  of atmospheric
32    deposition of metals (US  EPA, 2009a, section 9 A.23}. For example, the presence or absence of
33    a specific species of lichen can be used as a bioindicator of metal or organics contamination.
34    PBDEs detected in moss and lichens in Antarctica indicate long-range transport of PM
35    components (Yogui and Sericano 2008). In the U.S. Blue Ridge Mountains, a study linked metal
36    concentrations in mosses to elevation and tree canopy species at some sites but not with
                                               5-17
      Draft Do Not Quote or Cite                                             March 2010

-------
 1    concentrations of metals in the O horizon of soil (Schilling, 2002). A limitation to employing
 2    mosses and lichens to detect for the presence of air pollutants is the difference in uptake
 3    efficiencies of metals between species. The European Moss Biomonitoring Network has been
 4    shown to be useful in Europe for estimating general trends in metal concentrations and
 5    identification of some sources of trace contaminants, however, quantification of ecological
 6    effects is not possible due to the variability of species responses (US EPA, 2009a, section
 7    9.4.2.3).
 8          A potentially important regulating ecosystem service of plants is their capacity to
 9    sequester contaminants (US EPA, 2009a, section 9.4.5.3). Ongoing research on the application
10    of plants to environmental remediation efforts are yielding some success in removing heavy
11    metals and organics from contaminated sites (phytoremediation) with tolerant plants such as the
12    willow tree (Salix spp.) and members of the family Brassicaceae (US EPA, 2009a, section
13    9.4.5.3). Tree canopies can be used in urban locations to capture particulates and improve air
14    quality (Freer-Smith et al., 2004). Plant foliage is a sink for Hg and other metals and this
15    regulating ecosystem service may be impacted by atmospheric deposition of trace metals.
16          An ecological endpoint (phytochelatin concentration) associated with presence of metals
17    in the environment has been correlated with the ecological effect of tree mortality (Grantz et al.,
18    2003). Metal stress may be contributing to tree injury and forest decline in the Northeastern U.S.
19    where red spruce populations are declining with increasing elevation. Quantitative assessment of
20    PM damage to forests potentially could be conducted by overlaying PM sampling data and
21    elevated phytochelatin levels. However, limited data on phytochelatin levels in other species
22    currently hinders use of this peptide as a general biomarker for PM.
23          The presence of PM in the atmosphere affects ambient radiation as discussed in the ISA
24    which can impact the amount of sunlight received by plants (US EPA, 2009a,  section 9.4.4).
25    Atmospheric PM can change the radiation reaching leaf surfaces through attenuation and by
26    converting direct radiation to diffuse radiation. Diffuse radiation is more uniformly distributed in
27    a tree canopy, allowing radiation to reach lower leaves. The net effect of PM on photosynthesis
28    depends on the reduction of photosynthetically active radiation (PAR) and the increase in the
29    diffuse fraction of PAR. Decreases in crop yields (provisioning ecosystem service) have been
30    attributed to regional scale air pollution, however, global models suggest that the diffuse light
31    fraction  of PAR can increase growth (US EPA, 2009a, section 9.4.4).
32    Soil and Nutrient Cycling
33          Many of the major indirect plant responses to PM deposition are chiefly soil-mediated
34    and depend on the chemical composition of individual components of deposited PM.  Major
35    ecosystem services impacted by PM deposition to soils include support services such as nutrient
36    cycling,  products such as crops and regulating flooding and water quality.  Upon entering the soil
                                                5-18
      Draft Do Not Quote or Cite                                            March 2010

-------
 1    environment, PM pollutants can alter ecological processes of energy flow and nutrient cycling,
 2    inhibit nutrient uptake to plants, change microbial community structure and, affect biodiversity.
 3    Accumulation of heavy metals in soils depends on factors such as local soil characteristics,
 4    geologic origin of parent soils, and metal bioavailability.  It can be difficult to assess the extent
 5    to which observed heavy metal concentrations in soil are of anthropogenic origin (US EPA,
 6    2009a, section 9.4.5.1). Trace element concentrations are higher in some soils that are remote
 7    from air pollution sources due to parent material and local geomorphology.
 8          Heavy metals such as Zn, Cu, and Cd and  some pesticides can interfere with
 9    microorganisms that are responsible for decomposition of soil litter, an important regulating
10    ecosystem service that serves as a source of soil nutrients (US EPA, 2009a, sections 9.4.5.1 and
11    9.4.5.2). Surface litter decomposition is reduced in soils having high metal concentrations.  Soil
12    communities have associated bacteria, fungi, and invertebrates that are essential to soil nutrient
13    cycling processes. Changes to the relative species abundance and community composition can
14    be quantified to measure impacts of deposited PM to soil biota.  A mutualistic relationship exists
15    in the rhizophere (plant root zone) between plant roots, fungi, and microbes. Fungi in
16    association with plant roots form mycorrhizae that are essential for nutrient uptake by plants.
17    The role of mychorrizal fungi in plant uptake of metals from soils and effects of deposited PM
18    on soil microbes is discussed in section 9.4.5.2 of the ISA.
19    Wildlife
20          Animals play  a significant role in ecosystem function including nutrient cycling and crop
21    production (supporting ecosystem service), and as a source of food (provisioning ecosystem
22    service). Cultural ecosystem services provided by wildlife include bird and animal watching,
23    recreational hunting and fishing.  Impacts on these services are dependent upon the
24    bioavailability of deposited metals and organics and their respective toxicities to ecosystem
25    receptors.  Pathways of PM exposure to fauna include ingestion, absorption and trophic transfer.
26    Bioindicator species (known as sentinel organisms) can provide evidence of contamination due
27    to atmospheric pollutants. Use of sentinel species can be of particular value because chemical
28    constituents of deposited PM are difficult to characterize and have varying bioavailability (US
29    EPA, 2009a, section 9.4.5.5). Snails readily bioaccumulate contaminants such as PAHs and
30    trace metals. These organisms have been deployed as biomonitors for urban pollution and have
31    quantifiable biomarkers of exposure including growth inhibition, impairment of reproduction,
32    peroxidomal proliferation and induction of metal detoxifying proteins (metallothioneins)
33    (Gomet-de Vaufleury, 2000; Regoli, 2006). Earthworms have also been used as sensitive
34    indicators  of soil metal contamination.
35          Evidence of deposited PM effects on animals is limited (US EPA, 2009a, section 9.4.5.5).
36    Trophic transfer of pollutants of atmospheric origin has been demonstrated in limited studies.
                                                5-19
      Draft Do Not Quote or Cite                                              March 2010

-------
 1   PM may also be transferred between aquatic and terrestrial compartments.  There is limited
 2   evidence for biomagnifications of heavy metals up the food chain except for Hg which is well
 3   known to move readily through environmental compartments (US EPA, 2009a section 9.4.5.6).
 4   Bioconcentration of POPs and PBDEs in the Arctic and deep-water oceanic food webs indicates
 5   the global transport of particle-associated organics (US EPA, 2009a, section 9.4.6). Salmon
 6   migrations are contributing to metal accumulation in inland aquatic systems, potentially
 7   impacting the provisioning and cultural ecosystem service of fishing (US EPA, 2009a, section
 8   9.4.6). Stable isotope analysis can be applied to establish linkages between PM exposure and
 9   impacts to food webs, however, the use of this evaluation tool is limited for this ecological
10   endpoint due to the complexity of most trophic interactions (US EPA 2009a, section 9.4.5.6).
11   Foraging cattle have been used to assess atmospheric deposition and subsequent bioaccumulation
12   of Hg and trace metals and their impacts on provisioning services (US EPA, 2009a, section
13   9.4.2.3).
14   Water
15          New limited information on impacts of deposited PM on receiving water bodies indicate
16   that the ecosystem services of primary production, provision of fresh water, regulation of climate
17   and floods, recreational fishing and water purification are adversely impacted by atmospheric
18   inputs of metals and organics (US EPA, 2009a, sections 9.4.2.3 and 9.4.5.4).  Deposition of PM
19   to surfaces in urban settings increases the metal and organic component of storm water runoff
20   (US EPA, 2009a, sections 9.4.2.3). This atmospherically-associated pollutant burden can then be
21   toxic to aquatic biota.
22          Atmospheric deposition can be the primary source of some organics and metals to
23   watersheds. The contribution of atmospherically deposited PAHs to aquatic food webs was
24   demonstrated in high elevation mountain lakes with no other anthropogenic contaminant sources
25   (US EPA, 2009a, section 9.4.6). Metals associated with PM deposition limit phytoplankton
26   growth, impacting aquatic trophic structure.  Long-range atmospheric transport of 47 pesticides
27   and degradation products to the snowpack in seven national parks in the Western U.S. was
28   recently quantified indicating PM-associated contaminant  inputs to receiving waters during
29   spring snowmelt (Hageman et al., 2006).
30         •   What new techniques are available to improve our understanding of ecosystem
31            effects associated with metal and organic components of PM?
32   Regionally-based comprehensive ecological studies
33          The recently completed Western Airborne Contaminants Assessment Project (WACAP)
34   is the most comprehensive database on contaminant transport and PM depositional effects on
35   sensitive ecosystems in the U.S. In this project, the transport, fate, and ecological impacts of

                                               5-20
     Draft Do Not Quote or Cite                                            March 2010

-------
 1    anthropogenic contaminants from atmospheric sources were assessed from 2002 to 2007 in seven
 2    ecosystem components  (air, snow, water, sediment, lichen, conifer needles and fish) in eight
 3    core national parks (Landers et al., 2008). The goals of the study were to identify where the
 4    pollutants were accumulating, identify ecological indicators for those pollutants causing
 5    ecological harm, and to  determine the source of the air masses most likely to have transported
 6    the contaminants to the parks (US EPA, 2009a, section 9.4.6). Collected data were analyzed to
 7    identify probable local, regional and/or global sources of deposited PM components and their
 8    concurrent effects  on ecological receptors. The study concluded that bioaccumulation of semi-
 9    volatile organic compounds (SOCs) was observed throughout park ecosystems (Landers et al.,
10    2008). Findings from this study included the observation of an elevational gradient in PM
11    deposition with greater accumulation at higher altitude areas of the parks.  Furthermore, specific
12    ecological indicators were indentified in the WACAP that can be useful in assessing
13    contamination on larger spatial scales.  For example, quantification of concentrations of selected
14    pesticides in second-year conifer needles served  as a method for regional-scale comparison of
15    pollutant distribution (Landers et al., 2008).
16          In the WACAP study, bioaccumulation and biomagnification of airborne contaminants
17    were demonstrated on a regional scale in remote ecosystems in the Western United States.
18    Contaminants were shown to accumulate geographically based on proximity to individual
19    sources or source areas, primarily agriculture and industry (Landers et al.,  2008). This finding
20    was counter to the original working hypothesis that most of the contaminants found in western
21    parks would originate from eastern Europe and Asia  (Landers et al., 2008  p 6-8). The WACAP
22    study represents an experimental design in which ecological effects could  be correlated to
23    ambient pollutant levels on a regional scale. Although this assessment focuses on chemical
24    species that are components of PM, it does not specifically assess the effects of particulates
25    versus gas-phase forms; therefore, in most cases it is difficult to apply the  results to this
26    assessment based on paniculate concentration and size fraction (US EPA,  2009a, section 9.4.6).
27    There is a need for ecological modeling of PM components in different environmental
28    compartments to further elucidate links between PM and ecological indicators.
29          Europe and other countries are using the  critical load approach to assess pollutant effects
30    at the level of the ecosystem. This type of assessment requires site-specific data and information
31    on individual species responses to PM.  In respect to trace metals and organics, there are
32    insufficient data for the vast majority of U.S. ecosystems to calculate critical loads, however, a
33    methodology is being presented in the NOx/SOx Secondary REA (US EPA 2009h) to calculate
34    atmospheric concentrations from deposition that may be applicable to other environmental
35    contaminants.

                                               5-21
      Draft Do Not Quote or Cite                                            March 2010

-------
 1         •  Is there currently available information on ambient levels of PM that cause
 2            adverse effects on ecosystem components?
 3          As reviewed above, there is considerable data on impacts of PM on ecological receptors,
 4    but few studies that link ambient PM levels to observed effect. This is due, in part, to the nature,
 5    deposition, transport and fate of PM in ecosystems. PM is not a single pollutant, but a
 6    heterogeneous mixture of particles differing in size, origin and chemical composition (US EPA,
 7    2009a, section 9.4.1). The heterogeneity of PM exists not only within individual particles or
 8    samples from individual sites, but to even a greater extent, between samples from different sites.
 9    Since vegetation and other ecosystem components are affected more by particulate chemistry
10    than size fraction, exposure to a given mass concentration of airborne PM may lead to widely
11    differing plant or ecosystem responses, depending on the particular mix of deposited particles.
12           Many of the PM components bioaccumulate over time in organisms or plants making
13    correlations to ambient levels of PM impossible.  For example, in the WACAP study, SOC
14    accumulation in vegetation and air showed different patterns, possibly because each medium
15    absorbs different types of SOCs with varying efficiencies (Landers et al., 2008).
16          Bioindicator organisms demonstrated biological effects including growth inhibition,
17    metallothionein induction and reproductive impairment when exposed to complex mixtures of
18    ambient air pollutants (US EPA 2009a, section 9.4.5.5). Other studies quantify uptake of metals
19    and organics by plants or animals. However, due to the difficulty in correlating individual PM
20    components to a specific physiological response, these studies are limited. Furthermore, there
21    may be differences in uptake between species  such as differing responses to metal uptake
22    observed in mosses and lichens (US EPA 2009a, section 9.4.2.3). PM may also biomagnify
23    across trophic levels confounding efforts to link atmospheric concentrations to physiological
24    endpoints (US EPA, 2009a, section 9.4.5.6).
25          Evidence of PM effects that are linked to a specific ecological endpoint can be observed
26    when ambient levels are exceeded. Most direct ecosystem effects associated with particulate
27    pollution occur in severely polluted areas near industrial point sources (quarries, cement kilns,
28    metal smelting) (US EPA, 2009a, sections 9.4.3 and 9.4.5.7) . Extensive research on biota near
29    point sources provide some of the best evidence of ecosystem function impacts and demonstrates
30    that deposited PM has the potential to alter species composition over long time scales.
31    Ecological field studies conducted in proximity to Cu-Ni smelter in Harjavalta, Finland indicated
32    ecological structure and community composition are altered in response to PM and these effects
33    decrease with increasing distance from the point source (US EPA 2009a, section 9.4.5.7). The
34    ISA indicates at 4 km distance, species composition of vegetation, insects, birds, and soil
35    microbiota changed, and within 1 km only the most resistant organisms were surviving (US
36    EPA, 2009a, section 9.4.5.7). Heavy metal concentrations were quantified in understory plant
                                                5-22
      Draft Do Not Quote or Cite                                              March 2010

-------
 1    species growing at varying distance from the Harjavalta smelter (Salemaa et al., 2004). Heavy
 2    metal concentrations were highest in bryophytes, followed by lichens and were lowest in
 3    vascular plants. At the Harjavalta smelter there are clear links between PM deposition levels,
 4    ecological endpoints and compromised ecosystem structure.  However, these conditions are not
 5    reflective of ambient concentrations of PM in the majority of US ecosystems (US EPA, 2009a,
 6    section 9.4.7).

 7    5.3.3   Preliminary Staff Conclusions
 8         •   A number of significant environmental effects that either have already occurred or are
 9             currently occurring are linked to deposition of chemical constituents found in ambient
10             PM.
11         •   Ecosystem services can be adversely impacted by PM in the environment, including
12             supporting, provisioning, regulating and cultural services.
13         •   The lack of sufficient information to relate specific ambient concentrations of
14             particulate metals and organics to a degree of impairment of a specific ecological
15             endpoint hinders our ability to identify a range of appropriate indicators, levels, forms
16             and averaging times of a distinct secondary standard to protect against associated
17             effects.
18         •   Data from regionally-based ecological studies can be used to establish probable local,
19             regional and/or global sources of deposited PM  components and their concurrent
20             effects on  ecological receptors.
21           Collectively  taking into consideration the responses to specific questions regarding the
22    adequacy of the current secondary PM standards for ecological effects, we revisit the
23    overarching question: "does available scientific information, as reflected in the ISA, support or
24    call into question the adequacy of the protection for ecosystems afforded by the current suite of
25    secondary PM standards?" Staff reaches the preliminary conclusion that the available
26    information is insufficient to assess the adequacy of the protection for ecosystems afforded by
27    the current suite of PM secondary standards. Ecosystem effects linked to PM are  difficult to
28    determine because the changes may not be observed until pollutant deposition has occurred for
29    many decades. Because the high levels necessary to cause injury occur only near a few limited
30    point sources and/or on a very local scale, protection against these effects  alone may not provide
31    sufficient basis for considering a separate secondary NAAQS based on the ecological effects of
32    particulate metals and  organics. Data on ecological responses clearly linked with atmospheric
33    PM is not abundant  enough to perform a quantitative analysis although the WACAP study may
34    represent an opportunity for quantification at a regional scale. At this time, we conclude that
35    available evidence is not sufficient for establishing a distinct national standard for ambient PM
36    based on ecosystem effects of particulates not addressed in the NOx/SOx secondary review (e.g.
37    metals, organics).
                                                5-23
      Draft Do Not Quote or Cite                                              March 2010

-------
 1           Staff considered the appropriateness of continuing to use the PM2.5 and PMio size
 2    fractions as the indicators for protection of ecological effects of PM.  Though the chemical
 3    constitution of individual particles can be strongly correlated with size, the relationship between
 4    particle size and particle composition can also be quite complex, making it difficult in most cases
 5    to use particle size as a surrogate for chemistry.  At this time it remains to be determined as to
 6    what extent PM secondary standards focused on a given size fraction would result in reductions
 7    of the ecologically relevant constituents of PM for any  given area.

 8    5.3.4   Key Uncertainties and Areas for Future Research and Data Collection
 9           The above discussions identify linkages between ecological effects of deposited PM and
10    potential impacts to ecosystem services.  Unfortunately, our ability to relate ambient
11    concentrations of PM to  ecosystem response is hampered by a number of significant data gaps
12    and uncertainties. These limitations include the presence of multiple ecological stressors
13    confounding attempts to link specific ecosystem responses to PM deposition.  These stressors
14    can be anthropogenic (e.g. habitat destruction, eutrophication, other pollutants) or natural (e.g.
15    drought, fire, disease). Deposited PM interacts with other stressors to affect ecosystem patterns
16    and processes. Furthermore, the environmental effects of deposited PM are decoupled in space
17    and time from the point of emission confounding efforts to identify ecological perturbations
18    attributed to PM deposition.
19           A second source  of uncertainty lies in predicting the amount of PM deposited to sensitive
20    receptors from measured concentrations of PM in the ambient air.  This makes it difficult to
21    relate a given air concentration to a receptor response, an important factor in being able to set a
22    national ambient air quality standard. A multitude of factors such as the mode of deposition
23    (wet, dry and occult), wind speed, surface roughness or stickiness, elevation, particle
24    characteristics (e.g. size, shape, chemical composition), and relative humidity exert varying
25    degrees of influence on the deposition velocities for different PM components in any point in
26    time.  Composition of ambient PM varies in time and space and the particulate mixture may have
27    synergistic, antagonistic  or additive effects on ecological receptors depending upon the chemical
28    species present.  Furthermore, presence of co-occurring pollutants make it difficult to attribute
29    observed effects to ecological receptors to PM alone  or one component of deposited PM.
30           Third, each ecosystem has developed within a context framed by the topography,
31    underlying bedrock, soils, climate, meteorology, hydrologic regime, natural and land use history,
32    and species composition that make it unique from all others. Sensitivity of ecosystem response
33    is highly variable in space and time.  Because of this variety and lack of sufficient baseline data
34    on each of these features for most ecosystems, it is currently not possible to extrapolate with
35    confidence any effect from one ecosystem to another.

                                                 5-24
      Draft Do Not Quote or Cite                                              March 2010

-------
 1    5.4    MATERIALS
 2    5.4.1  Scope
 3          Welfare effects on materials associated with deposition of PM include both physical
 4    damage (materials damage effects) and impaired aesthetic qualities (soiling effects). Because the
 5    effects of PM are exacerbated by the presence of acidic gases and can be additive or synergistic
 6    due to the complex mixture of pollutants in the air and surface characteristics of the material, this
 7    discussion will also include those particles and gases that are associated with the presence of
 8    ambient NOx and SOx, as well as NHa and NHx for completeness. Building upon the
 9    information presented in the last Staff Paper (US EPA, 2005), and including the limited new
10    information presented in Chapter 9 of the PM ISA (US EPA, 2009a) and Annex E. Effects of
11    NOy, NHx, and SOx on Structures and Materials of the Integrated Science Assessment for
12    Oxides of Nitrogen and Sulfur-Ecological Criteria (NOx/SOx ISA) (US EPA, 2008c) the
13    following sections consider the policy-relevant aspects of physical damage and aesthetic soiling
14    effects of PM on materials including metal and stone.
15          The ISA concludes that evidence is sufficient to support a causal relationship between
16    PM and effects on materials (US EPA, 2009a, sections 2.5.4 and 9.5.4). The deposition of PM
17    can physically affect materials, adding to the effects of natural weathering processes, by
18    potentially promoting or accelerating the corrosion of metals, by degrading paints and by
19    deteriorating building materials such as stone, concrete and marble (US EPA, 2009a, section
20    9.5).  Particles contribute to these physical effects because of their electrolytic,  hygroscopic and
21    acidic properties, and their ability to sorb corrosive gases (principally 802). In  addition, the
22    deposition of ambient PM can reduce the aesthetic appeal of buildings and objects through
23    soiling. Particles consisting primarily of carbonaceous compounds cause soiling of commonly
24    used building materials and culturally important items such as statues and works of art. Soiling
25    is the deposition of particles on surfaces by impingement, and the accumulation of particles on
26    the surface of an exposed material results in degradation of its appearance (US EPA 2009a,
27    section 9.5).  Soiling can be remedied by cleaning  or washing, and depending on the  soiled
28    material, repainting.

29    5.4.2  Adequacy of the Current Standard
30          In considering the  adequacy of the suite of secondary standards, staff addresses the
31    following overarching question:
32
33
34
 Does available scientific information, as reflected in the ISA support or call into question
the adequacy of the protection for materials afforded by the current suite of secondary PM
                                      standards?
                                                5-25
     Draft Do Not Quote or Cite                                              March 2010

-------
 1          To inform the answer to this overarching question, staff has posed a specific question to
 2    aid in assessing the available scientific evidence as related to materials damage and soiling
 3    attributed to PM deposition as presented in the ISA (US EPA, 2009a).
 4         •  What new evidence is available to improve our understanding of effects of PM on
 5            materials and linking ambient concentrations to materials damage?
 6          The majority of available new studies on materials effects of PM are from outside the
 7    U.S., however, they provide limited new data for consideration of the secondary standard.
 8          Metal and stone are susceptible to damage by ambient PM. Considerable research has
 9    been conducted on the  effects of air pollutants on metal surfaces due to the economic importance
10    of these materials, especially steel, zinc, aluminum, and copper. Chapter 9 of the PM ISA and
11    Annex E of the NOx/SOx ISA summarize the results of a number of studies on the corrosion of
12    metals (US EPA, 2009a; US EPA, 2008c).  Moisture is the single greatest factor promoting metal
13    corrosion, however, deposited PM can have additive, antagonistic or synergistic effects.  In
14    general, SC>2 is more corrosive than NOx although mixtures of NOx, SC>2 and other particulate
15    matter corrode some metals at a faster rate than either pollutant alone (US EPA, 2008c, Annex
16    E.5.2).  Information from both the PM ISA and NOx/SOx ISA suggest that the extent of damage
17    to metals due to ambient PM is variable and dependent upon the type of metal, prevailing
18    environmental conditions, rate of natural weathering and presence or absence of other pollutants.
19          The PM ISA and NOx/SOx ISA summarize the results of a number of studies on PM and
20    stone surfaces. While it is clear from the available information that gaseous air pollutants, in
21    particular SO2, will promote the deterioration of some types of stones under specific conditions,
22    carbonaceous particles (non-carbonate carbon) and particles containing metal oxides may help to
23    promote the decay process.  Studies on  metal and stone summarized in the ISA do not show an
24    association between particle size, chemical composition and frequency of repair.
25          A limited number of new studies available on materials damage effects of PM since the
26    last review consider the relationship between pollutants and biodeterioration of structures
27    associated with microbial communities  that colonize monuments and buildings (US EPA 2009a,
28    section 9.5). Presence of air pollutants may synergistically enhance microbial deterioration
29    processes.  The role of heterotrophic bacteria, fungi and cyanobacteria  in biodeterioration varied
30    by local meterological  conditions and pollutant components. In a comparative study of
31    biodeterioration processes on monuments in Latin America, limestone  deterioration  at the Mayan
32    site of Uxmal was enhanced by biosolubilization by metabolic acids from bacteria and fungi
33    while destruction of the Cathedral of La Plata was attributed primarily  to atmospheric pollutants
34    (Herrera and Videla, 2006).
35          PM deposition onto surfaces such as metal, glass,  stone and paint can lead to soiling.
36    Soiling results when PM accumulates on an object and  alters the optical characteristics
                                               5-26
      Draft Do Not Quote or Cite                                             March 2010

-------
 1    (appearance). The reflectivity of a surface may be changed or presence of particulates may alter
 2    light transmission.  These effects can impact the aesthetic value of a structure or result in
 3    reversible or irreversible damage to statues, artwork and architecturally or culturally significant
 4    buildings. Due to soiling of building surfaces by PM, the frequency and duration of cleaning
 5    may be increased. Soiling affects the aesthetic appeal of painted surfaces. In addition to natural
 6    factors, exposure to PM may give painted surfaces a dirty appearance.  Pigments in works of art
 7    can be degraded or discolored by atmospheric pollutants, especially sulfates (US EPA, 2008c,
 8    Annex E-15).
 9          Formation of black crusts due to carbonaceous compounds and buildup of microbial
10    biofilms results in discoloration of surfaces.  Black crust includes a carbonate component derived
11    from building material and organic carbon (OC) and elemental carbon  (EC). In limited new
12    studies quantifying the OC and EC contribution to soiling by black crust, OC predominated over
13    EC at almost all locations (Bonazza et al., 2005).  Limited new studies suggest that traffic is the
14    major source of carbon associated with black crust formation (Putaud,  2004) and that soiling of
15    structures in Oxford, UK showed a relationship with traffic and NO2 concentrations (Viles and
16    Gorbushina, 2003).  These findings attempt to link atmospheric concentrations of PM to
17    observed damage. However, no data on rates of damage are available and all studies were
18    conducted outside of the U.S.

19    5.4.3  Preliminary Staff Conclusions
20          Available evidence in regards to materials damage and soiling supports the following
21    observations:
22         •   Materials damage and soiling that occur through natural weathering processes are
23             enhanced by exposure to atmospheric pollutants, most notably SO2 and  particulate
24             sulfates.
25         •   While ambient particles play a role in the corrosion of metals and in  the weathering of
26             materials, no quantitative relationships between ambient particle concentrations and
27             rates of damage have been established.
28         •   While soiling associated with fine and course particles can result in increased cleaning
29             frequency and repainting of surfaces, no quantitative relationships between particle
30             characteristics and the frequency of cleaning or repainting have been established.
31         •   Limited new data on the role of microbial colonizers in biodeterioration processes and
32             contributions of black crust to soiling are not sufficient for quantitative analysis.
33         •   While several studies in the PM ISA  and NOx/SOx ISA suggest that particles can
34             promote corrosion of metals there remains insufficient evidence to relate corrosive
35             effects to specific particulate levels or to establish a quantitative relationship between
36             ambient PM and metal degradation. With respect to damage to calcareous stone,

                                                5-27
      Draft Do Not Quote or Cite                                             March 2010

-------
 1              numerous studies suggest that wet or dry deposition of particles and dry deposition of
 2              gypsum particles can enhance natural weathering processes.
 3           Revisiting the overarching policy question as to whether the available scientific evidence
 4    supports or calls into question the adequacy of the protection for materials afforded by the
 5    current suite of secondary PM standards, we reach the preliminary conclusion that no new
 6    evidence in this review calls into question the adequacy of the protection for materials afforded
 7    by the current standard. PM effects on materials can play no quantitative role in considering
 8    whether any revisions of the secondary PM NAAQS are appropriate at this time. However, in
 9    the absence of information that provides a basis for establishing a different level of control,
10    observations continue to support retaining an appropriate degree of control on both fine and
11    coarse particles to help address materials damage and soiling associated with PM.

12    5.4.4  Key Uncertainties and Areas for Future Research and Data Collection
13           Quantitative relationships are needed between particle size, concentration, chemical
14    concentrations and frequency of repainting  and repair.  Deposition rates of airborne PM to
15    surfaces would provide an indication of rate and degree of damage to surfaces.  There is
16    considerable uncertainty with regard to interaction of co-pollutants in regards to materials
17    damage and soiling processes.

18    5.5    REFERENCES
19    Audet P; Charest C (2007). Heavy metal phytoremediation from a meta-analytical perspective. Environ Pollut, 147:
20           231-237.
21    Bonazza A; Sabbioni C; Ghedini N. (2005). Quantitative data on carbon fractions in interpretation of black crusts
22           and soiling on European built heritage. Atmos Environ, 39: 2607-2618.
23    Bond TC; Streets DG; Yarber KF; Nelson SM; Woo J-H; Klimont Z. (2004). A technology-based global inventory
24           of black and organic carbon emissions from combustion. J Geophys Res, 109.
25    Bond TC; Sun H. (2005). Can reducing black carbon emissions counteract global warming?. Environ Sci Technol,
26           39: 5921-5926.
27    CCSP (2009). Atmospheric Aerosol Properties and Climate Impacts, A Report by the U.S. Climate Change Science
28           Program and the Subcommittee on Global Change Research. [Mian Chin,  Ralph A. Kahn, and Stephen E.
29           Schwartz (eds.)]. National Aeronautics and Space Administration, Washington, D.C., USA.
30    Forster P; Ramaswamy V; Artaxo P; Betss R; Fahey DW; Hay wood J; Lean J; Lowe DC; Myhre G; Nganga J; Prinn
31           R; Raga G;Schultz M; Van Borland R. (2007). Changes in atmospheric constituents and in radiative
32           forcing. In Solomon, S.; Qin,D.; Manning,  M.; Chen, Z.; Marquis, M.; Averyt, K. B.; Tignor, M.; Miller,
33           H. L. (Ed.),Climate Change 2007: the physical science basis. Contribution of Working Group I to the fourth
3 4           assessment report of the intergovernmental panel.
3 5    Freer-Smith PH; El-khatib A; Taylor G. (2004). Capture of paniculate pollution by trees: a comparison of species
3 6           typical of semi-arid areas (Ficus nitida and Eucalyptus globulus) with European and North American
37           species. Water Air Soil Pollut, 155: 173-187.
                                                   5-28
      Draft Do Not Quote or Cite                                                March 2010

-------
  1    Gitay H; Brown S; Easterling W; Jallow B. (2001). Ecosystems and their goods and services. In Climate change
  2            2001: impacts,adaptation and vulnerability. Contribution of working group II to the third assessment report
  3            of the Intergovernmental Panel on Climate Change (pp. 237-342). Cambridge, United Kingdom:
  4            Cambridge University Press.

  5    Givati A; Rosenfeld D. (2004). Quantifying precipitation suppression due to air pollution., 43: 1038-1056.

  6    Gomot-De Vaufleury A; Pihan F. (2002). Methods for toxicity assessment of contaminated soil by oral or dermal
  7            uptake in land snails: Metal bioavailability and bioaccumulation. Environ Toxicol Chem, 21: 820-827.

  8    Grantz DA; Garner JHB; Johnson DW. (2003). Ecological effects of paniculate matter. Environ Int, 29: 213-239.

  9    Hageman KJ; Simonich SL; Campbell DH; Wilson GR; Landers DH. (2006). Atmospheric deposition of current-use
10            and historic-use pesticides in snow at national parks in the western United States. Environ Sci Technol, 40:
11            3174-3180

12    Hassan R; Scholes R; Ash N. (2005). Ecosystems and human well-being: current state and trends, volume 1. Herrera
13            LK; Videla HA. (2004). The importance of atmospheric effects on biodeterioration of cultural heritage
14            constructional materials. Int Biodeterior Biodegradation,  54: 125-134. IPCC (2007). Summary for
15            Policymakers. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to
16            the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (Solomon, S; Quin D,
17            Manning M, Cehn Z; Marquis M; Averyt KB; Tignor M; and Miller HL (eds.) Cambridge University Press,
18            Cambridge United Kingdom and New York, NY, USA.

19    Herrera LK; Videla HA. (2004). The importance of atmospheric effects on biodeterioration of cultural heritage
20            constructional materials. Int Biodeterior Biodegradation,  54: 125-134.

21    IPCC (2007): Summary for Policymakers. In: Climate Change 2007: The Physical Science Basis. Contribution of
22            Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change
23            [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)].
24            Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

25    Jacobson MZ. (2000). A physically-based treatment of elemental carbon optics: implications for global direct
26            forcing of aerosols. Geophys Res Lett, 27: 217-220.

27    JacobsonMZ; Kaufman YJ. (2006). Wind reduction by aerosol particles. Geophys Res Lett, 33.

28    Kucera T; Horakova H; Sonska A (2008). Toxic metal ions in photoautotrophic organisms. Photosynthetica, 46:
29            481-489.

30    Landers DH; Simonich SL; Jaffe DA; Geiser LH; Campbell DH; Schwindt AR; Schreck CB; Kent ML; Hafner WD;
31            Taylor HE; Hageman KJ; Usenko S; Ackerman LK; Schrlau JE; Rose NL; Blett TF; Erway MM. (2008).
32            The Fate, Transport and Ecological Impacts of Airborne Contaminants in Western National Parks (USA).
33            U.S. Environmental Protection Agency, Office of Research and Development, NHEERL, Western Ecology
3 4            Division. Corvallis, Oregon. EPA/600/R-07/13 8

35    National Research Council. (2001). Climate change science: an analysis of some key questions. National Research
36            Council. National Academy Press. Washington, DC.

37    Parrish ZD; White JC; Isleyen M; Gent MPN; lannucci-Berger W; Eitzer BD; Kelsey JW; Mattina MI. (2006).
3 8            Accumulation of weathered polycyclic aromatic hydrocarbons (PAHs) by plant and earthworm species.
39            Chemosphere, 64: 609-618.
                                                      5-29
      Draft Do Not Quote or Cite                                                   March 2010

-------
  1    Patra M; Bhowmik N; Bandopadhyay B; Sharma A (2004). Comparison of mercury, lead and arsenic with respect to
  2            genotoxic effects on plant systems and the development of genetic tolerance. Environ Exp Bot, 52: 199-
  3            223.

  4    Putaud J-P; Raes F; Van Dengenen R; Bruggemann E; Facchini M-C; Decesari S; Fuzzi S; Gehrig R; Huglin C; Laj
  5            P; Lorbeer G; Maenhaut W; Mihalopoulos N; Muller K; Querol X; Rodriguez S; Schneider J; Spindler G;
  6            ten Brink H; Torseth K; Wiedensohler A.  (2004). A European aerosol phenomenology~2: chemical
  7            characteristics of paniculate matter at kerbside, urban, rural and background sites in Europe. Atmos
  8            Environ, 38: 2579-2595.

  9    Regoli F; Gorbi S; Fattorini D; Tedesco S; Notti A; Machella N; Bocchetti R; Benedetti M; Piva F. (2006). Use of
10            the land snail Helix aspersa sentinel organism for monitoring ecotoxicologic effects of urban pollution: An
11            integrated approach.,  114:63-69.

12    Salemaa M; Derome J; Helmisaari HS;  Nieminen T; Vanha-Majamaa I. (2004). Element accumulation in boreal
13            bryophytes,lichens and vascular plants exposed to heavy metal and sulfur deposition in Finland. Sci Total
14            Environ, 324: 141-160.

15    Sato M; Hansen J; Koch D; Lucis A; Ruedy R; Dubovik O; Holben B; Chin M; Novakov T. (2003). Global
16            atmospheric black carbon inferred from AAEONET. Presented at Proceedings of the National Academy of
17            Science.

18    Schilling JS; Lehman ME. (2002). Bioindication of atmospheric heavy metal deposition in the Southeastern US
19            using the moss Thuidium delicatulum. Atmos Environ, 36: 1611-1618.

20    Smith WH (1990). Forest nutrient cycling: toxic ions. In Air pollution and forests: interactions between air
21            contaminants and forest ecosystemsNew York, NY: Springer-Verlag.

22    Strydom C; Robinson C; Pretorius E; Whitcutt JM; Marx J; Bornman MS (2006). The effect of selected metals on
23            the central metabolic pathways in biology: A review. WaterSA, 32: 543-554.

24    U.S. EPA (1996). Air quality criteria for paniculate matter. U.S. Environmental Protection Agency. Research
25            Triangle Park, NC. EPA/600/P-95/00laF-cF.

26    US EPA (2004). Air Quality Criteria for Particulate Matter. National Center for Environmental Assessment, Office
27            of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711;
28            report no. EPA/600/P-99/002aF and EPA/600/P-99/002bF. October 2004.  Available:
29            http://www.epa.gov/ttn/naaqs/standards/pm/s_pm cr cd.html

30    US EPA (2005). Review of the National Ambient Air Quality Standards for Particulate Matter: Policy Assessment
31            of Scientific and Technical Information, OAQPS Staff Paper. Research Triangle Park, NC 27711: Office
32            of Air Quality Planning and Standards; report no. EPA EPA-452/R-05-005a. December 2005.
33                    Available: http://www.epa.gov/ttn/naaqs/standards/pm/s_pm cr  sp.html

34    U.S. EPA. (2006). Air Quality Criteria for Lead Final Report. U.S. Environmental Protection Agency, Washington,
3 5            DC, EPA/600/R-05/144aF-bF, 2006.

36    US EPA (2008a).  Integrated Review Plan for the National Ambient Air Quality Standards for Particulate Matter.
37            National Center for Environmental Assessment and Office of Air Quality Planning and Standards, U.S.
3 8            Environmental Protection Agency, Research Triangle Park, NC. Report No. EPA 452/R-08-004. March
39            2008. Available at:
40            http://www.epa.gov/ttn/naaqs/standards/pm/data/2008  03  final integrated review  plaapdf
                                                      5-30
      Draft Do Not Quote or Cite                                                   March 2010

-------
 1    US EPA (2008c). U.S. EPA. Integrated Science Assessment (ISA) for Oxides of Nitrogen and Sulfur Ecological
 2            Criteria (Final Report). U.S. Environmental Protection Agency, Washington, D.C., EPA/600/R-08/082F,
 3            2008.

 4    US EPA (2009a). U.S. EPA. Integrated Science Assessment for Paniculate Matter (Final Report). US.
 5            Environmental Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009.

 6    US EPA (2009h) Risk and Exposure Assessment for Review of the Secondary National Ambient Air Quality
 7            Standards for Oxides of Nitrogen and Oxides of Sulfur. (Final Report). US Environmental Protection
 8            Agency, Research Triangle Park, NC, EPA-452/R-09-008a

 9    US EPA (2010c) Policy Assessment for the Review of the Secondary National Ambient Air Quality Standards for
10            NOx and SOx, First External Review Draft. US Environmental Protection Agency, Research Triangle Park,
11            NC, EPA-452/P-10-006

12    Viles HA; Gorbushina AA. (2003). Soiling and microbial colonisation on urban roadside limestone: A three year
13            study in Oxford, England. ,38: 1217-1224.

14    Yogui G; Sericano J. (2008). Polybrominated diphenyl ether flame retardants in lichens and mosses from King
15            George Island, maritime Antarctica. Chemosphere, 73: 1589-1593.
                                                      5-31
      Draft Do Not Quote or Cite                                                   March 2010

-------
United States                          Office of Air Quality Planning and Standards          Publication No. EPA-452/P-10-003
Environmental Protection               Health and Environmental Impacts Division                               March 2010
Agency                                      Research Triangle Park, NC

-------