&EPA
United States
Environmental Protection
Agency
Third External Review Draft of
Air Quality Criteria for
Particulate Matter (April, 2002)
Volume I
-------
EPA/600/P-99/002aC
April 2002
Third External Review Draft
Air Quality Criteria for Participate Matter
Volume I
National Center for Environmental Assessment-RTP Office
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
1 DISCLAIMER
2
3 This document is an external review draft for review purposes only and does not constitute
4 U.S. Environmental Protection Agency policy. Mention of trade names or commercial products
5 does not constitute endorsement or recommendation for use.
6
7
8 PREFACE
9
10 National Ambient Air Quality Standards (NAAQS) are promulgated by the United States
11 Environmental Protection Agency (EPA) to meet requirements set forth in Sections 108 and 109
12 of the U.S. Clean Air Act (CAA). Sections 108 and 109 require the EPA Administrator (1) to list
13 widespread air pollutants that reasonably may be expected to endanger public health or welfare;
14 (2) to issue air quality criteria for them that assess the latest available scientific information on
15 nature and effects of ambient exposure to them; (3) to set "primary" NAAQS to protect human
16 health with adequate margin of safety and to set "secondary" NAAQS to protect against welfare
17 effects (e.g., effects on vegetation, ecosystems, visibility, climate, manmade materials, etc.); and
18 (5) to periodically (every 5 years) review and revise, as appropriate, the criteria and NAAQS for
19 a given listed pollutant or class of pollutants.
20 The original NAAQS for particulate matter (PM), issued in 1971 as "total suspended
21 parti culate" (TSP) standards, were revised in 1987 to focus on protecting against human health
22 effects associated with exposure to ambient PM less than 10 microns (< 10 //m) that are capable
23 of being deposited in thoracic (tracheobronchial and alveolar) portions of the lower respiratory
24 tract. Later periodic reevaluation of newly available scientific information, as presented in the
25 last previous version of this "Air Quality Criteria for Parti culate Matter" document published in
26 1996, provided key scientific bases for PM NAAQS decisions published in July 1997. More
27 specifically, the PM10 NAAQS set in 1987 (150 //g/m3, 24-h; 50 //g/m3, annual average) were
28 retained in modified form and new standards (65 //g/m3, 24-h; 15 //g/m3, annual average) for
29 particles <2.5 //m (PM25) were promulgated in July 1997.
April 2002 I-ii DRAFT-DO NOT QUOTE OR CITE
-------
1 This Third External Review Draft of revised Air Quality Criteria for Particulate Matter
2 assesses new scientific information that has become available mainly between early 1996 through
3 December 2001. The present draft is being released for public comment and review by the Clean
4 Air Scientific Advisory Committee (CASAC) to obtain comments on the organization and
5 structure of the document, the issues addressed, the approaches employed in assessing and
6 interpreting the newly available information on PM exposures and effects, and the key findings
7 and conclusions arrived at as a consequence of this assessment. Public comments and CASAC
8 review recommendations will be taken into account in making any appropriate further revisions
9 to this document for incorporation into a final draft. Evaluations contained in the present
10 document will be drawn on to provide inputs to associated PM Staff Paper analyses prepared by
11 EPA's Office of Air Quality Planning and Standards (OAQPS) to pose alternatives for
12 consideration by the EPA Administrator with regard to proposal and, ultimately, promulgation of
13 decisions on potential retention or revision of the current PM NAAQS.
14 Preparation of this document was coordinated by staff of EPA's National Center for
15 Environmental Assessment in Research Triangle Park (NCEA-RTP). NCEA-RTP scientific
16 staff, together with experts from other EPA/ORD laboratories and academia, contributed to
17 writing of document chapters; and earlier drafts of this document were reviewed by experts from
18 federal and state government agencies, academia, industry, and NGO's for use by EPA in support
19 of decision making on potential public health and environmental risks of ambient PM. The
20 document describes the nature, sources, distribution, measurement, and concentrations of PM in
21 outdoor (ambient) and indoor environments. It also evaluates the latest data on human exposures
22 to ambient PM and consequent health effects in exposed human populations (to support decision
23 making regarding primary, health-related PM NAAQS). The document also evaluates ambient
24 PM environmental effects on vegetation and ecosystems, visibility, and man-made materials, as
25 well as atmospheric PM effects on climate change processes associated with alterations in
26 atmospheric transmission of solar radiation or its reflectance from the Earth's surface or
27 atmosphere (to support decision making on secondary PM NAAQS).
28 The NCEA of EPA acknowledges the contributions provided by authors, contributors, and
29 reviewers and the diligence of its staff and contractors in the preparation of this document.
30
April 2002 I-iii DRAFT-DO NOT QUOTE OR CITE
-------
Air Quality Criteria for Particulate Matter
VOLUME I
EXECUTIVE SUMMARY E-l
1. INTRODUCTION 1-1
2. PHYSICS, CHEMISTRY, AND MEASUREMENT OF
PARTICULATE MATTER 2-1
3. CONCENTRATIONS, SOURCES, AND EMISSIGNS OF
ATMOSPHERIC PARTICULATE MATTER 3-1
Appendix 3 A: Spatial and Temporal Variability of the Nationwide
AIRS PM25 and PM10.25 Data Sets 3A-1
Appendix 3B: Aerosol Composition Data from the Speciation
Network 3B-1
Appendix 3C: Organic Composition of Particulate Matter 3C-1
Appendix 3D: Composition of Particulate Matter Source Emissions .... 3D-1
4. ENVIRONMENTAL EFFECTS OF PARTICULATE MATTER 4-1
5. HUMAN EXPOSURE TO PARTICULATE MATTER AND ITS
CONSTITUENTS 5-1
April 2002 I-iv DRAFT-DO NOT QUOTE OR CITE
-------
Air Quality Criteria for Particulate Matter
(cont'd)
VOLUME II
6. DOSIMETRY OF PARTICULATE MATTER 6-1
7. TOXICOLOGY OF PARTICULATE MATTER IN HUMANS AND
LABORATORY ANIMALS 7-1
8. EPIDEMIOLOGY OF HUMAN HEALTH EFFECTS FROM
AMBIENT PARTICULATE MATTER8-1 8-1
Appendix 8A: Short-Term PM Exposure-Mortality Studies:
Summary Table 8A-1
Appendix 8B: Particulate Matter-Morbidity Studies:
Summary Tables 8B-1
9. INTEGRATIVE SYNTHESIS 9-1
Appendix 9A: Key Quantitative Estimates of Relative Risk for
Particulate Matter-Related Health Effects Based on
Epidemiologic Studies of U.S. and Canadian Cities
Assessed in the 1996 Particulate Matter Air Quality
Criteria Document 9A-1
April 2002 I-v DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
Page
List of Tables I-xii
List of Figures I-xviii
Authors, Contributors, and Reviewers I-xxx
U.S. Environmental Protection Agency Project Team for Development of
Air Quality Criteria for Particulate Matter I-xli
Abbreviations and Acronyms I-xliv
EXECUTIVE SUMMARY E-l
1. INTRODUCTION 1-1
1.1 LEGISLATIVE REQUIREMENTS 1-1
1.2 HISTORY OF PREVIOUS PM CRITERIA AND NAAQS REVIEWS 1-2
1.2.1 The 1997 PMNAAQS Revision 1-3
1.2.2 Coordinated Particulate Matter Research Program 1-5
1.3 CURRENT PM CRITERIA AND NAAQS REVIEW 1-10
1.3.1 Key Milestones 1-10
1.3.2 Methods and Procedures for Document Preparation 1-12
1.3.3 Approach 1-13
1.3.4 Key Human Health Issues of Concern 1-14
1.4 DOCUMENT ORGANIZATION AND CONTENT 1-17
REFERENCES 1-19
2. PHYSICS, CHEMISTRY, AND MEASUREMENT OF PARTICULATE MATTER ... 2-1
2.1 PHYSICS AND CHEMISTRY OF PARTICULATE MATTER 2-2
2.1.1 Basic Concepts 2-2
2.1.2 Physical Properties and Processes 2-3
2.1.2.1 Definitions of Particle Diameter 2-3
2.1.2.2 Aerosol Size Distributions 2-4
2.1.2.3 Nuclei-Mode Particles 2-16
2.1.3 Chemistry of Atmospheric Particulate Matter 2-20
2.1.3.1 Chemical Composition and Its Dependence on Particle Size ... 2-20
2.1.3.2 Primary and Secondary Particulate Matter 2-21
2.1.3.3 Particle-Vapor Partitioning 2-22
2.1.3.4 Atmospheric Lifetimes and Removal Processes 2-25
2.1.4 Summary 2-28
2.2 MEASUREMENT OF PARTICULATE MATTER 2-28
2.2.1 Particle Measurements of Interest 2-28
2.2.2 Issues in Measurement of Particulate Matter 2-31
2.2.2.1 Treatment of Semivolatile Components of Particulate Matter .. 2-32
2.2.2.2 Upper Cut Point 2-34
2.2.2.3 Cut Point for Separation of Fine-Mode and Coarse-Mode
Particulate Matter 2-35
April 2002 I-vi DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
2.2.2.4 Treatment of Pressure, Temperature, and Relative Humidity ... 2-38
2.2.2.5 Time Resolution 2-39
2.2.2.6 Accuracy and Precision 2-39
2.2.3 Measurement of Semivolatile Particulate Matter 2-41
2.2.3.1 Particulate Nitrates 2-43
2.2.3.2 Semivolatile Organic Compounds 2-46
2.2.3.3 Particle-Bound Water 2-59
2.2.4 U. S. Environmental Protection Agency Monitoring Methods 2-65
2.2.4.1 The Federal Reference Methods for Measurement of
Equilibrated Mass for PM10, PM2 5, and PM10_2 5 2-65
2.2.5 Speciation Monitoring 2-78
2.2.6 Inorganic Elemental Analyses 2-80
2.2.6.1 Energy Dispersive X-ray Fluorescence (EDXRF) 2-81
2.2.6.2 Synchrotron Induced X-ray Fluorescence (S-XRF) 2-82
2.2.6.3 Proton Induced X-ray Emission (PIXE) 2-82
2.2.6.4 Proton Elastic Scattering Analysis (PESA) 2-83
2.2.6.5 Total Reflection X-ray Fluorescence (TRXRF) 2-83
2.2.6.6 Instrumental Neutron Activation Analysis (TNAA) 2-84
2.2.6.7 Atomic Absorption Spectrophotometry (AAS) 2-85
2.2.6.8 Inductively Coupled Plasma with Atomic Emission
Spectroscopy (ICP-AES) 2-86
2.2.6.9 Inductively Coupled Plasma with Mass Spectroscopy
(ICP-MS) 2-86
2.2.6.10 Scanning Electron Microscopy (SEM) 2-87
2.2.7 Elemental and Organic Carbon in Particulate Matter 2-88
2.2.8 Ionic Species 2-98
2.2.9 Continuous Monitoring 2-98
2.2.9.1 Continuous Measurement of Mass 2-99
2.2.9.2 Continuous Measurement of Elemental and Organic
Carbon 2-105
2.2.9.3 Continuous Measurements of Nitrate and Sulfate 2-107
2.2.9.4 Continuous Ion Chromatography of Water-Soluble Ions 2-109
2.2.9.5 Measurements of Individual Particles 2-109
2.2.9.6 Determination of Aerosol Surface Area in Real Time 2-111
2.2.9.7 Light Scattering 2-112
2.2.10 Low Flow Filter Samples for Multiday Collection of Particulate
Matter 2-112
2.3 SUMMARY 2-114
2.3.1 Atmospheric Physics and Chemistry of Particles 2-114
2.3.2 Measurement of Atmospheric Particles 2-117
REFERENCES 2-121
April 2002 I-vii DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
3. CONCENTRATIONS, SOURCES, AND EMISSIONS OF ATMOSPHERIC
PARTICULATE MATTER 3-1
3.1 INTRODUCTION 3-1
3.2 PATTERNS AND TRENDS IN AMBIENT PM CONCENTRATIONS 3-2
3.2.1 Seasonal Variability in PM Concentrations 3-13
3.2.2 Diurnal (Circadian) Variability in PM Concentrations 3-27
3.2.3 Relations Among Particulate Matter in Different Size Fractions 3-29
3.2.4 Relations Between Mass and Chemical Component Concentrations 3-31
3.2.5 Spatial Variability in Particulate Matter and its Components 3-37
3.3 SOURCES OF PRIMARY AND SECONDARY PARTICULATE
MATTER 3-56
3.3.1 Chemistry of Secondary PM Formation 3-59
3.3.2 The Long-Range Transport of Particulate Matter from Outside the
United States 3-68
3.3.3 Source Contributions to Ambient PM Determined by Receptor
Models 3-73
3.3.4 Emissions Estimates for Primary Particulate Matter, and Precursors
to Secondary Particulate Matter (SO2, NOX, VOCs, and NH3) in the
United States 3-82
3.3.5 Uncertainties of Emissions Inventories 3-88
3.4 SUMMARY AND CONCLUSIONS 3-91
REFERENCES 3-95
APPENDIX 3 A: Spatial and Temporal Variability of the Nationwide AIRS
PM25 andPM10.25 Data Sets 3A-1
REFERENCES 3A-35
APPENDIX 3B: Aerosol Composition Data From The Speciation Network 3B-1
REFERENCES 3B-36
APPENDIX 3C: Organic Composition of Particulate Matter 3C-1
REFERENCES 3C-14
APPENDIX 3D: Composition of Particulate Matter Source Emissions 3D-1
REFERENCES 3D-26
4. ENVIRONMENTAL EFFECTS OF PARTICULATE MATTER 4-1
4.1 INTRODUCTION 4-1
4.2 IMPACTS ON VEGETATION AND ECOSYSTEMS 4-1
4.2.1 Particle Deposition 4-2
4.2.1.1 Wet Deposition 4-3
April 2002 I-viii DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
4.2.1.2 Dry Deposition 4-5
4.2.1.3 Occult Deposition 4-30
4.2.1.4 Magnitude of Deposition 4-33
4.2.2 Effects on Vegetation and Ecosystems 4-44
4.2.2.1 Direct Effects of Particulate Matter on Individual Plant
Species 4-46
4.2.2.2 Indirect Effects of Parti culate Matter on Natural Ecosystems ... 4-63
4.2.3 Urban Ecosystems 4-133
4.2.4 Ecosystem Goods and Services and Their Economic Valuation 4-137
4.3 EFFECTS ON VISIBILITY 4-141
4.3.1 Introduction 4-141
4.3.2 Factors Affecting Atmospheric Visibility 4-143
4.3.2.1 Optical Properties of the Atmosphere and Atmospheric
Particles 4-143
4.3.2.2 Relative Humidity Effects on Particle Size and Light-
Scattering Properties 4-149
4.3.3 Relationships Between Particles and Visibility 4-152
4.3.4 Photographic Modeling of Visibility Impairment 4-156
4.3.5 Visibility Monitoring Methods and Networks 4-158
4.3.6 Visibility Impairment: Trends and Current Conditions 4-161
4.4 EFFECTS ON MATERIALS 4-165
4.4.1 Corrosive Effects of Particles and Sulfur Dioxide on Man-Made
Surfaces 4-166
4.4.1.1 Metals 4-166
4.4.1.2 Painted Finishes 4-168
4.4.1.3 Stone and Concrete 4-171
4.4.2 Soiling and Discoloration of Man-Made Surfaces 4-176
4.4.2.1 Stones and Concrete 4-177
4.4.2.2 Household and Industrial Paints 4-177
4.5 EFFECTS OF ATMOSPHERIC PARTICULATE MATTER ON GLOBAL
CLIMATE CHANGE PROCESSES AND THEIR POTENTIAL HUMAN
HEALTH AND ENVIRONMENTAL IMPACTS 4-178
4.5.1 Solar Ultraviolet Radiation Transmission Impacts on Human Health
and the Environment: Atmospheric Parti culate Matter Effects 4-180
4.5.1.1 Potential Effects of Increased Ultraviolet Radiation
Transmission 4-180
4.5.1.2 Effects of Airborne Particles on Transmission of Solar
Ultraviolet Radiation Through the Atmosphere 4-184
4.5.2 Global Warming Processes, Human Health and Environmental
Impacts, and Roles of Atmospheric Particle 4-188
April 2002 I-ix DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
4.5.2.1 Bases for Concern Regarding Global Warming and Climate
Change 4-188
4.5.2.2 Airborne Particle Relationships to Global Warming and
Climate Change 4-195
4.6 SUMMARY 4-200
4.6.1 Particulate Matter Effects on Vegetation and Ecosystems 4-200
4.6.2 Particulate Matter-Related Effects on Visibility 4-208
4.6.3 Particulate Matter-Related Effects on Materials 4-209
4.6.4 Effects of Atmospheric Particulate Matter on the Transmission of
Solar Ultraviolet Radiation and Global Warming Processes 4-210
REFERENCES 4-212
APPENDIX 4A: Colloquial and Latin Names 4A-1
5. HUMAN EXPOSURE TO PARTICULATE MATTER AND ITS
CONSTITUENTS 5-1
5.1 INTRODUCTION 5-1
5.1.1 Purpose 5-1
5.1.2 Particulate Matter Mass and Constituents 5-2
5.1.3 Relationship to Past Documents 5-3
5.2 STRUCTURE FOR THE CHAPTER 5-4
5.3 BASIC CONCEPTS OF EXPOSURE 5-5
5.3.1 Components of Exposure 5-5
5.3.2 Methods To Estimate Personal Exposure 5-7
5.3.2.1 Direct Measurement Methods 5-7
5 3.2.2 Indirect Methods (Modeling Methods) 5-8
5.3.2.3 Methods for Estimating Personal Exposure to Ambient
Particulate Matter 5-15
5.4 SUMMARY OF PARTICULATE MATTER MASS DATA 5-20
5.4.1 Types of Particulate Matter Exposure Measurement Studies 5-20
5.4.2 Available Data 5-21
5.4.2.1 Personal Exposure Data 5-21
5.4.2.2 Microenvironmental Data 5-25
5.4.2.3 Reanalyses of Previously-Reported Particulate Matter
Exposure Data 5-31
5.4.3 Factors Influencing and Key Findings on Particulate Matter
Exposures 5-34
5.4.3.1 Relationship of Personal/Microenvironmental Particulate
Matter with Ambient Particulate Matter 5-34
5.4.3.2 Factors That Affect Relationship between Personal
Exposure and Ambient PM 5-53
April 2002 I-x DRAFT-DO NOT QUOTE OR CITE
-------
Table of Contents
(cont'd)
5.4.3.3 Impact of Ambient Sources on Exposures to Particulate
Matter 5-70
5.4.3.4 Correlations of Particulate Matter with Other Pollutants 5-72
5.5 SUMMARY OF PARTICULATE MATTER CONSTITUENT DATA 5-76
5.5.1 Introduction 5-76
5.5.2 Monitoring Studies That Address Particulate Matter Constituents 5-77
5.5.3 Key Findings 5-77
5.5.3.1 Correlations of Personal and Indoor Concentrations with
Ambient Concentrations of Particulate Matter Constituents .... 5-77
5.5.4 Factors Affecting Correlations Between Ambient Measurements and
Personal or Microenvironmental Measurements of Particulate Matter
Constituents 5-85
5.5.5 Limitations of Available Data 5-86
5.6 IMPLICATIONS OF USING AMBIENT PARTICULATE MATTER
CONCENTRATIONS IN EPIDEMIOLOGIC STUDIES OF PARTICULATE
MATTER HEALTH EFFECTS 5-87
5.6.1 Potential Sources of Error Resulting from Using Ambient Particulate
Matter Concentrations in Epidemiologic Analyses 5-87
5.6.2 Associations Between Personal Exposures and Ambient Particulate
Matter Concentrations 5-89
5.6.3 Role of Compositional Differences in Exposure Characterization for
Epidemiology 5-93
5.6.4 Role of Spatial Variability in Exposure Characterization for
Epidemiology 5-94
5.6.5 Analysis of Exposure Measurement Error Issues in Particulate Matter
Epidemiology 5-95
5.6.5.1 Analysis of Exposure Measurement Errors in Time-Series
Studies 5-96
5.6.5.2 Analysis of Exposure Measurement Errors in Long-Term
Epidemiology Studies 5-99
5.6.5.3 Conclusions from Analysis of Exposure Measurement
Errors on Particulate Matter Epidemiology 5-101
5.7 SUMMARY OF KEY FINDINGS AND LIMITATIONS 5-102
REFERENCES 5-110
April 2002 I-xi DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
Number Page
1-1 Key Scientific Uncertainties Related to the Source-To-Response Framework 1-8
1-2 Research Topics and Questions Recommended by National Research Council
to be Addressed by Expanded Particulate Matter Research Program 1-9
1 -3 Schedule for Development of the Current Revised Parti culate Matter Air
Quality Criteria Document 1-11
2-1 Comparison of Ambient Particles Fine Mode and Coarse Mode 2-29
2-2 Parti culate Matter Components/Parameters of Interest for Health, Ecological,
or Radiative Effects; for Source Category Apportionment Studies; or for Air
Quality Modeling Evaluation Studies 2-30
2-3 Summary of Sensitivity Studies of WINS Impactor Performance 2-70
2-4 PM2 5 Samplers Currently Designated as FRMs for PM2 5 Mass Concentrations . . . 2-71
3-1 Distribution of Ratios of PM2 5 to Pm10 and Correlations Between PM2 5 and
PM10, PM2 5 and PM10_2 5, and PM10_2 5 and PM10 Found at Collocated Monitoring
Sites in Seven Aerosol Characteristic (EPA/HEI) Regions in 1999 3-30
3-2 Concentrations (in ng/m3) of PM25, PM10_25, and Selected Elements (ng/m3)
in the PM2 5 and PM10_2 5 Size Ranges With Standard Deviations and
Correlations Between Elements and PM2 5 Mass in Philadelphia, PA 3-32
3-3 Concentrations (in ng/m3) of PM25, PM10_25 and Selected Elements in the
PM2 5 and PM10_2 5 Size Range With Standard Deviations and Correlations (r)
Between Elements and PM25 and PM10_2 5 Mass in Phoenix, AZ 3-33
3-4 Measures of the Spatial Variability of PM25 Concentrations Within Selected
Metropolitan Statistical Areas 3-38
3-5 Measures of the Spatial Variability of PM10_25 Concentrations Within Selected
Metropolitan Statistical Areas 3-46
3-6 Correlation Coefficients for Spatial Variation of PM25 Mass and Different
Sources for Pairs of Sampling Sites in the South Coast Air Basin (1986) 3-53
3-7 Correlation Coefficients for Spatial Variation of PM25 Mass and Different
Components for Pairs of Sampling Sites in Philadelphia (1994) 3-54
April 2002 I-xii DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number
3-8 Constituents of Atmospheric Particles and Their Major Sources 3-57
3-9 Receptor Model Source Contributions to PM2 5 3-77
3-10 Receptor Model Source Contributions to PM10 3-78
3-11 Emissions of Primary PM25 by Various Sources in 1999 3-84
3-12 Emissions of Precursors to Secondary PM25 Formation by Various Sources
in 1999 3-85
3B-1 Summary Statistics for PM25 Concentrations During February Through June 2000
Obtained by Collocated Federal Reference Method Samplers (in //g/m3) 3B-2
3B-2 Summary Statistics for the Speciation Samplers at Bismarck, ND (in //g/m3) 3B-3
3B-3 Summary Statistics for the Speciation Samplers at Boston, MA (in //g/m3) 3B-6
3B-4 Summary Statistics for the Speciation Samplers at Bronx Botanical Garden, NY
(in Mg/m3) 3B-8
3B-5 Summary Statistics for the Speciation Samplers at Chicago, IL (in //g/m3) 3B-11
3B-6 Summary Statistics for the Speciation Samplers at Fresno, CA (in //g/m3) 3B-14
3B-7 Summary Statistics for the Speciation Samplers at Houston, TX (in //g/m3) 3B-16
3B-8 Summary Statistics for the Speciation Samplers at Lewis, FL (in //g/m3) 3B-19
3B-9 Summary Statistics for the Speciation Samplers at Philadelphia, PA (in //g/m3) . . 3B-21
3B-10 Summary Statistics for the Speciation Samplers at Phoenix, AZ (in //g/m3) 3B-24
3B-11 Summary Statistics for the Speciation Samplers at Portland, OR (in //g/m3) 3B-26
3B-12 Summary Statistics for the Speciation Samplers at Salt Lake City, UT
(in //g/m3) 3B-28
3B-13 Summary Statistics for the Speciation Samplers at St. Louis, MO (in //g/m3) .... 3B-30
3B-14 Summary Statistics for the Speciation Samplers at Seattle, WA (in //g/m3) 3B-33
April 2002 I-xiii DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number
3C-1 Particulate Organic and Elemental Carbon Concentrations (in //g C/m3) Based
on Studies Published After 1995 3C-2
3C-2 Particulate Organic Compound Concentrations (in ng C/m3) Based on Studies
Published after 1990 at Selected Sites 3C-5
3D-1 Average Abundances of Major Elements in Soil and Crustal Rock 3D-2
3D-2 Summary of Particle-Phase Organic Compounds Present in Fine Particle
Road Dust Sample 3D-6
3D-3 Composition of Fine Particles Released by Various Stationary Sources in the
Philadelphia Area 3D-7
3D-4a Organic and Elemental Carbon Fractions of Diesel and Gasoline Engine
Particulate Matter Exhaust 3D-13
3D-4b Contribution of Organic Carbon to Particulate Matter Carbon Emissions
in Motor Vehicle Exhaust Collected from Vehicles Operated on Chassis
Dynamometers 3D-14
3D-5 Emission Rates (mg/mi) for Constituents of Particulate Matter from
Gasoline and Diesel Vehicles 3D-15
3D-6 Summary of Particle-Phase Organic Compounds Emitted from Motor
Vehicles 3D-17
3D-7 Mass Emissions, Organic Carbon, and Elemental Carbon Emissions from
Residential Combustion of Wood 3D-20
3D-8 Summary of Particle-Phase Organic Compounds Emitted from the
Combustion of Wood in Fireplaces 3D-21
3D-9 Mean Aerosol Composition at Tropical Site (Sriwijaya University, Sumatra,
Indonesia) Affected Heavily by Biomass Burning Emissions 3D-23
4-1 Relative Importance of Wet, Dry, Particulate, and Total Deposition to Three
Forest Sites 4-4
4-2 Key Determinants of Dry Particulate Deposition to Vegetation 4-7
April 2002 I-xiv DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number
4-3 Reported Mean Deposition Velocities (Vd) for Sulfate, Chlorine, Nitrate, and
Ammonium and Ion-Containing Particles 4-28
4-4 Representative Empirical Measurements of Deposition Velocity (Vd) for
Particulate Deposition 4-29
4-5 Reported Mean Deposition Velocities for Potassium, Sodium, Calcium, and
Magnesium Base Cation Containing Particles 4-29
4-6 Relative Magnitudes of Wet, Dry, and Occult Deposition of Nitrates (NO3") and
Sulfates (SO4"2) to Three Forest Sites Subject to Similar Gas- and Liquid-Phase
Pollutant Concentrations During Spring and Summer 4-33
4-7 Mean Annual Nitrogen Deposition (equivalents/ha/yr) from Fine and Coarse
Particles Compared to Total Nitrogen Deposition from all Sources to a Variety
of Forest Ecosystems 4-36
4-8 Mean Annual Sulfate Deposition (equivalents/ha/yr) from Fine and Coarse
Particles Compared to Total Sulfur Deposition from all Sources to a Variety
of Forest Ecosystems 4-37
4-9 Mean Annual Base Cation Deposition (equivalents/ha/yr) from Fine and Coarse
Particles Compared to Total Base Cation Deposition from all Sources to a Variety
of Forest Ecosystems 4-38
4-10 Mean (±SE) Particle Size, Deposition Rates, and Derived Deposition Velocities
(Vd) for Heavy Metal Deposition to the Upper Canopy (Inert Plates or Leaves)
of an Upland Oak Forest 4-41
4-11 Total Heavy Metal Deposition to Temperate Latitude Forests 4-42
4-12 Annual Bulk Deposition Input of Seven Heavy Metals to the Hubbard Brook
Experimental Forest (43 ° 56' N Latitude, 71 ° 45' W Longitude), White Mountain
National Forest, NH, for the Period 1975 to 1991 (grams per hectare) 4-43
4-13 Ecosystem Services 4-65
4-14 Ecosystem Functions Impacted by Air Pollution Effects on Temperate Forest
Ecosystems 4-71
4-15 Types of Plant Responses to Ultraviolet-B Radiation 4-84
April 2002 I-xv DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number
4-16 Nitrogen-Saturated Forests in North America, Including Estimated N Inputs
and Outputs 4-88
4-17 Primary Goods and Services Provided by Ecosystems 4-138
4-18 Corrosive Effects of Particulate Matter and Sulfur Dioxide on Metals 4-169
4-19 Corrosive Effects of Particulate Matter and Sulfur Dioxide on Stone 4-172
4-20 Examples of Impacts Resulting from Projected Changes in Extreme Climate
Events 4-192
5-1 Classes of PM Exposure and Concentration Definitions 5-6
5-2 Activity Pattern Studies Included in the Consolidated Human Activity
Database 5-11
5-3 Personal Exposure Models for PM 5-13
5-4 Summary of Recent Personal Exposure Studies 5-22
5-5 Summary of Recent Microenvironmental Measurement Studies 5-26
5-6 Papers Interpreting PM Exposure Studies 5-32
5-7 Personal Monitoring Studies for PM: Measured Concentrations and
Correlation Coefficients 5-37
5-8 Mean Concentration for PM Mass Reported for the Baltimore (Williams et al.,
2000a,b,c) and the Fresno (Evans et al., 2000) Studies 5-44
5-9 Daily-Average Correlation Coefficients Reported for the Baltimore
(Williams et al., 2000a,b,c) and the Fresno (Evans et al., 2000) Studies 5-45
5-10 Regression Analysis Reported for Indoor/Outdoor Relationships for the Baltimore
(Williams et al., 2000a,b,c) and the Fresno (Evans et al., 2000) Studies 5-45
5-11 Volume Mean Diameter and Maximum PM2 5 Concentrations of Indoor
Particle Sources 5-67
5-12 Correlations Between Personal PM2 5 and Ambient Pollutant Concentrations 5-74
April 2002 I-xvi DRAFT-DO NOT QUOTE OR CITE
-------
List of Tables
(cont'd)
Number
5-13 Correlations Between Hourly Personal PM2 5 and Gaseous Pollutants 5-76
5-14 Studies That Have Measured Particulate Matter Constituents in Personal
Exposure Samples 5-79
5-15 Studies That Have Measured Particulate Matter Constituents in
Microenvironmental Samples 5-80
5-16 Summary Statistics for Personal, Indoor, and Outdoor Concentrations of
Selected Aerosol Components in Two Pennsylvania Communities 5-82
5-17 Statistical Correlation of Outdoor (x) Versus Indoor (y) Concentration for
Measured Species 5-84
April 2002 I-xvii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
Number Page
1-1 A general framework for integrating particulate-matter research 1-7
2-1 Number of particles as a function of particle diameter: (a) number
concentrations are shown on a logarithmic scale to display the wide
range by site and size; (b) number concentrations for the average urban
distribution are shown on a linear scale for which the area under any part
of the curve is proportional to particle number in that size range 2-6
2-2 Particle volume distribution as a function of particle diameter: (a) for the
averaged rural and urban-influenced rural number distributions shown in
Figure 2-1 and a distribution from south central New Mexico, and (b) for
the averaged urban and freeway-influenced urban number distributions
shown in Figure 2-1 2-7
2-3 Distribution of coarse (c), accumulation (a), and nuclei or ultrafine (n),
mode particles by three characteristics, a) number (N), b) surface area (S)
and c) volume (V) for the grand average continental size distribution 2-8
2-4 Volume size distribution, measured in traffic, showing fine-mode and
coarse-mode particles and the nuclei and accumulation modes within
the fine particle mode 2-9
2-5 An idealized distribution of ambient particulate matter showing fine-mode
particles and coarse-mode particles and the fractions collected by size-selective
samplers 2-11
2-6 Specified particle penetration (size-cut curves) through an ideal
(no-particle-loss) inlet for five different size-selective sampling criteria 2-13
2-7 Comparison of penetration curves for two PM10 beta gauge samplers using
cyclone inlets 2-15
2-8 Particle growth curves showing fully reversible hygroscopic growth of
sulfuric acid (H2SO4) particles, deliquescent growth of ammonium sulfate
[(NH4)2 SO4] particles at about 80% relative humidity (RH), hygroscopic
growth of ammonium sulfate solution droplets at RH greater than 80%, and
hysteresis (the droplet remains supersaturated as the RH decreases below
80%) until the crystallization point is reached 2-25
2-9 Theoretical predictions and experimental measurements of growth of
NH4HSO4 and particles at relative humidity between 95 and 100% 2-26
April 2002 I-xviii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
2-10 Schematic showing major nonvolatile and semivolatile components of PM2 5 .... 2-33
2-11 Paniculate matter concentrations in Spokane, WA, during the August 30,
1996 dust storm 2-37
2-12 Amount of ammonium nitrate volatilized from Teflon filters, expressed as
a percentage of the measured PM2 5 mass, for the SCAQS and CalTech
studies, for spring and fall sampling periods 2-44
2-13 Aerosol water content expressed as a mass percentage, as a function of
relative humidity 2-63
2-14 Schematic diagram of the sample collection portion of the PM25 FRM
sampler 2-67
2-15 Schematic view of the final design of the WINS 2-68
2-16 Evaluation of the final version of the WINS 2-69
2-17 Schematic diagram showing the principle of virtual impaction 2-77
2-18 This thermogram, for a sample containing rock dust (a carbonate source)
and diesel exhaust, shows three traces that correspond to temperature, filter
transmittance, and FID detector response 2-93
2-19 Comparison of mass measurements with collocated RAMS (real-time data),
PC-BOSS (1-h data), FRM PM25 sampler (average of 24-h data), and a
conventional TEOM monitor (real-time data) 2-102
2-20 Size distribution of particles divided by chemical classification into organic,
marine, and crustal 2-110
3-la 1999-2000 county-wide average annual meanPM10 concentrations (//g/m3) 3-6
3-lb 1999-2000 highest county-wide 98th percentile 24-h average PM10 concentrations
(Mg/m3) 3-6
3-2 Nationwide trend in ambient PM10 concentration from 1989 through 1998 3-8
3-3 Trend in PM10 annual mean concentrations by EPA region, 1989-1998 3-8
April 2002 I-xix DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
3-4a 1999-2000 county-wide average annual meanPM25 concentrations (//g/m3) 3-9
3-4b 1999-2000 highest county-wide 98th percentile 24-h average PM2 5
concentrations (//g/m3) 3-9
3-5 Collection of annual distribution of 24-h average PM25 concentrations
observed in U.S. and Canadian health studies conducted during the 1980's
and early 1990's 3-10
3-6a 1999-2000 estimated county-wide average annual mean PM10_25
concentrations (//g/m3) 3-14
3-6b 1999-2000 estimated county-wide highest 98th percentile 24-h average PM10_2 5
concentrations (//g/m3) 3-14
3-7a,b Quarterly distribution of 24-h average PM25 concentrations for selected
monitors in the (a) Columbia, SC; (b) Detroit, MI; (c) Chicago, IL; and
(d) Los Angeles, CA MSAs. Values for the lowest, lower quartile, median,
upper quartile and highest concentrations are shown in the figures 3-16
3-7c,d Quarterly distribution of 24-h average PM25 concentrations for selected
monitors in the (a) Columbia, SC; (b) Detroit, MI; (c) Chicago, IL; and
(d) Los Angeles, CA MSAs. Values for the lowest, lower quartile, median,
upper quartile and highest concentrations are shown in the figures 3-17
3-8 Concentrations of PM25 and PM10 measured in the four MAACS cities 3-19
3-9a,b Quarterly distribution of 24-h average PM10_25 concentrations for selected
sites in the (a) Columbia, SC; (b) Detroit, MI; and (c) Los Angeles, CA
metropolitan statistical areas 3-20
3-9c Quarterly distribution of 24-h average PM10_25 concentrations for selected
sites in the (a) Columbia, SC; (b) Detroit, MI; and (c) Los Angeles, CA
metropolitan statistical areas 3-21
3-10 Frequency distribution of 24-h average PM25 concentrations measured at
the monitoring site at the Presbyterian home in southwestern Philadelphia
from 1992 to 1995 3-22
3-11 Concentrations of 24-h average PM2 5 measured at the EPA site in Phoenix,
AZ from 1995 to 1997 3-24
April 2002 I-xx DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
3-12 Frequency distribution of 24-h average PM25 concentrations measured at the
EPA site in Phoenix, AZ from 1995 to 1997 3-24
3-13 Frequency distribution of 24-h average PM25 measurements obtained from all
California Air Resources Board dichotomous sampler sites from 1989 to 1998 . . . 3-25
3-14 Frequency distribution of 24-h average PM10_25 concentrations obtained from all
California Air Resource Board Dichotomous sampler sites from 1989 to 1998 . . . 3-25
3-15 Concentrations of 24-h average PM25 measured at the Riverside-Rubidoux
site from 1989 to 1998 3-26
3-16 Frequency distribution of 24-h average PM25 concentrations measured at the
Riverside-Rubidoux site from 1989 to 1994 3-26
3-17 Intraday variability of hourly average PM25 concentrations across the
United States 3-28
3-18 Occurrence of differences between pairs of sites in three metropolitan
statistical areas 3-45
3-19 Intersite correlation coefficients for PM25, PM10, and PM10_25 3-49
3-20 PM2 5 chemical components in downtown Los Angeles and Burbank (1986)
have similar characteristics 3-51
3-21 Concentrations of PM2 5 chemical components in Rubidoux and downtown
Los Angeles (1986) 3-52
3-22 Monthly average Saharan dust components of the aerosol sampled in Miami,
FL, from 1974 to 1996 3-70
3-23 PM2 5 and PM10 concentrations measured at Chilliwack Airport, located in
northwestern Washington State, just before and during the Asian desert dust
episode of April and May 1998 3-71
3-24 Time series of 24-h average PM10 concentrations observed in the Rio Grande
Valley during May 1998 3-72
3-25 PM10 concentrations observed in St. Louis, MO, during May 1998 3-72
April 2002 I-xxi DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
3 A-l Philadelphia, PA-NJ MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-3
3 A-2 Washington, DC MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-4
3 A-3 Norfolk, VA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-5
3 A-4 Columbia, SC MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-6
3 A-5 Atlanta, GA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-7
3 A-6 Birmingham, AL MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-8
3 A-7 Tampa, FL MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-9
3 A-8 Cleveland, OH MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations;
(c) Intersite correlation statistics, for each data pair, the correlation
coefficient, (P90, coefficient of divergence) and number of measurements
are given 3A-10
April 2002 I-xxii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
3 A-9 Pittsburgh, PA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-11
3 A-10 Steubenville, OH-Weirton, WV MSA. (a) Locations of sampling sites by
AIRS ID#; (b) Quarterly distribution of 24-h average PM25 concentrations;
(c) Intersite correlation statistics, for each data pair, the correlation coefficient,
(P90, coefficient of divergence) and number of measurements are given 3A-12
3 A-l 1 Detroit MI MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations; (c) Intersite correlation
statistics, for each data pair, the correlation coefficient, (P90, coefficient of
divergence) and number of measurements are given 3A-13
3 A-12 Grand Rapids, MI MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-14
3 A-13 Milwaukee, WI MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3 A-l 5
3 A-14 Chicago, IL MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations; (c) Intersite correlation
statistics, for each data pair, the correlation coefficient, (P90, coefficient of
divergence) and number of measurements are given 3A-16
3 A-15 Gary, IN MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations; (c) Intersite correlation
statistics, for each data pair, the correlation coefficient, (P90, coefficient of
divergence) and number of measurements are given 3A-17
3 A-16 Louisville, KY MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3 A-l 8
April 2002 I-xxiii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
3 A-17 St. Louis, MO MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-19
3 A-18 Baton Rouge, LA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-20
3 A-19 Kansas City, KS-MO MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given 3A-21
3 A-20 Dallas, TX MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations; (c) Intersite correlation
statistics, for each data pair, the correlation coefficient, (P90, coefficient of
divergence) and number of measurements are given 3A-22
3 A-21 Boise, ID MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations; (c) Intersite correlation
statistics, for each data pair, the correlation coefficient, (P90, coefficient of
divergence) and number of measurements are given 3A-23
3A-22 Salt Lake City, UT MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c)
Intersite correlation statistics, for each data pair, the correlation coefficient,
(P90, coefficient of divergence) and number of measurements are given 3A-24
3 A-23 Seattle, WA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c)
Intersite correlation statistics, for each data pair, the correlation coefficient,
(P90, coefficient of divergence) and number of measurements are given 3A-25
3 A-24 Portland, OR MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c)
Intersite correlation statistics, for each data pair, the correlation coefficient,
(P90, coefficient of divergence) and number of measurements are given 3A-26
April 2002 I-xxiv DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
3 A-25 Los Angeles-Long Beach, CA MSA. (a) Locations of sampling sites by
AIRS ID; (b) Quarterly distribution of 24-h average PM25 concentrations;
(c) Intersite correlation statistics, for each data pair, the correlation coefficient,
(P90, coefficient of divergence) and number of measurements are given 3A-27
3 A-26 Riverside-San Bernadino, CA MSA. (a) Locations of sampling sites by
AIRS ID#; (b) Quarterly distribution of 24-h average PM25 concentrations;
(c) Intersite correlation statistics, for each data pair, the correlation coefficient,
(P90, coefficient of divergence) and number of measurements are given 3A-28
3 A-27 San Diego, CA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM25 concentrations; (c)
Intersite correlation statistics, for each data pair, the correlation coefficient,
(P90, coefficient of divergence) and number of measurements are given 3A-29
3 A-28 Columbia, SC MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_25 concentrations;
(c) Intersite correlation coefficients and number of measurements 3A-30
3 A-29 Detroit, MI MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM10_25 concentrations; (c) Intersite correlation
coefficients and number of measurements 3A-31
3 A-30 Cleveland, OH MSA. (a) Locations of sampling sites by AIRS ID#; (b)
Quarterly distribution of 24-h average PM10_25 concentrations; (c) Intersite
correlation coefficients and number of measurements 3A-32
3 A-31 Steubenville, OH-Weirton, WV MSA. (a) Locations of sampling sites by
AIRS ID#; (b) Quarterly distribution of 24-h average PM10_25 concentrations;
(c) Intersite correlation coefficients and number of measurements 3A-33
3 A-32 St. Louis, MO-IL MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_25 concentrations; (c) Intersite
correlation coefficients and number of measurements 3A-34
3D-1 Size distribution of particles generated in a laboratory resuspension chamber ... 3D-3
3D-2 Size distribution of California source emissions, 1986 3D-4
3D-3 Chemical abundances for PM2 5 emissions from paved road dust in
Denver, CO 3D-5
April 2002 I-xxv DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
3D-4 Chemical abundances for PM2 5 emissions from wood burning in
Denver, CO 3D-19
4-1 A simplified resistance catena representing the factors controlling deposition
of particles to the surface 4-8
4-2 The relationship between deposition velocity of selected paniculate materials and
the distribution of the material between the coarse- and fine-aerosol fractions 4-9
4-3 The relationship between particle diameter and deposition velocity for
particles 4-10
4-4 Vertical stratification of diverse, chemically speciated particles in a mixed
oak forest 4-24
4-5 The relationship between particle size and concentration below a spruce canopy
with wind velocity at a height of 16.8 m equaling 5ms"1 4-25
4-6 Mean (±SE) percent of total nitrogen, sulfur, or base cation deposition
contributed by fine plus coarse particles 4-39
4-7 Effects of environmental stress on forest trees are presented on a hierarchial
scale for the leaf, branch, tree, and stand levels of organization 4-67
4-8 Nitrogen cycle (dotted lines indicate processes altered by nitrogen saturation) .... 4-89
4-9 Diagrammatic overview of excess nitrogen (N) in North America 4-91
4-10 Schematic of sources and sinks of hydrogen ions in a forest 4-104
4-11 Calcium deposition in >2-//m particles, <2-//m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites 4-113
4-12 Magnesium deposition in >2-//m particles, <2-//m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites 4-114
4-13 Potassium deposition in >2-//m particles, <2-//m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites 4-115
4-14 Base cation deposition in >2-//m particles, <2-//m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites 4-116
April 2002 I-xxvi DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
4-15 Total cation leaching (total height of bar) balanced by sulfate and nitrate
estimated from particulate deposition (assuming no ecosystem retention,
paniculate sulfur and nitrogen) and by other sources (both deposition and
internal) of sulfate and nitrate (other sulfur and nitrogen sources) and by
other anions in the Integrated Forest Study sites 4-117
4-16 Soil exchangeable Ca+2 pools and net annual export of Ca+2 (deposition minus
leaching times 25 years) in the Integrated Forest Study sites 4-118
4-17 Soil exchangeable Mg+2 pools and net annual export of Mg+2 (deposition minus
leaching times 25 years) in the Integrated Forest Study sites 4-118
4-18 Soil exchangeable K2+ pools and net annual export of K2+ (deposition minus
leaching times 25 years) in the Integrated Forest Study sites 4-119
4-19a Simulated soil solution mineral acid anions and base cations in the red
spruce site with no change, 50% N and S deposition, and 50% base
cation deposition 4-122
4-19b Simulated soil solution mineral acid anions and base cations in the red
spruce site with no change, 50% N and S deposition, and 50% base
cation deposition 4-123
4-20a Simulated soil solution Al and soil base saturation in the red spruce site
with no change, 50% N and S deposition, and 50% base cation deposition 4-124
4-20b Simulated soil solution Al and soil base saturation in the red spruce site
with no change, 50% N and S deposition, and 50% base cation deposition 4-125
4-2la Simulated soil solution mineral acid anions and base cations in the
Coweeta site with no change, 50% N and S deposition, and 50% base
cation deposition 4-127
4-2Ib Simulated soil solution mineral acid anions and base cations in the
Coweeta site with no change, 50% N and S deposition, and 50% base
cation deposition 4-128
4-22 Relationship of plant nutrients and trace metals with vegetation 4-132
4-23 Linkages among various ecosystem goods and services (food, water,
biodiversity, forest products) and other driving forces (climate change) 4-139
April 2002 I-xxvii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
4-24 Light reflected from a target toward an observer 4-145
4-25 Light-scattering efficiency factor (per cross sectional area), Q, for a
homogeneous sphere with an index of refraction of 1.50 as a function
of the size parameter, a = 7iD/A 4-148
4-26 Particle growth curve as a function of relative humidity showing deliquescent
growth of ammonium sulfate [(NH4)2 SO4] particles at the deliquescent point
(A, about 80% relative humidity [RH]), reversible hygroscopic growth of
ammonium sulfate solution droplets at RH greater than 80%, and hysteresis
(the droplet remains supersaturated as the RH decreases below 80%) until the
crystallization point (B, about 38% RH) is reached 4-150
4-27 Comparison of extinction (Mm"1) and visual range (km) 4-154
4-28 Relative humidity adjustment factor, f(RH), for ammonium sulfate as a
function of relative humidity 4-156
4-29a Aggregate visibility trends (in deciviews) for 10 eastern Class 1 areas 4-163
4-29b Aggregate visibility trends (in deciviews) for 26 western Class 1 areas 4-163
4-3Oa Eastern class I area aggregate trends in aerosol light extinction on the 20%
haziest days, including trends by major aerosol component 4-164
4-3Ob Western class I area aggregate trends in aerosol light extinction on the 20%
haziest days, including trends by major aerosol component 4-164
4-31 Processes involved in stratospheric ozone depletion because of man's
production of CFCs, halons, and other trace gases are shown to the left 4-181
4-32 Estimated global mean radiative forcing exerted by gas and various particle
phase species for the year 2000, relative to 1750 4-199
5-1 Comparison of correlation coefficients for longitudinal analyses of personal
exposure versus ambient concentrations for individual subjects for PM2 5
and sulfate 5-48
5-2 Personal exposure versus ambient concentrations for PM2 5 and sulfate 5-50
April 2002 I-xxviii DRAFT-DO NOT QUOTE OR CITE
-------
List of Figures
(cont'd)
Number
5-3 Regression analyses of aspects of daytime personal exposure to PM10 estimated
using data from the PTEAM study 5-52
5-4 Air-exchange rates measured in homes throughout the United States 5-59
5-5 Box plots of hourly air-exchange rates stratified by season in Boston, MA,
during 1998 5-60
5-6 Geometric mean infiltration factor (indoor/outdoor ratio) for hourly nighttime,
nonsource data for two seasons 5-60
5-7 Comparison of deposition rates from this study with literature values (adapted
from Abt et al., 2000b) 5-62
5-8 Penetration efficiencies and deposition rates from models of nightly average
data 5-63
5-9 Mean hourly indoor/outdoor particle concentration ratio from an unoccupied
residence in Fresno, CA, during spring 1999 5-68
5-10 Personal versus outdoor SO4= in State College, PA 5-83
5-11 Plots of nonambient exposure to PM10, (a) daytime individual values from
PTEAM data and (b) daily-average values from THEES data 5-92
April 2002 I-xxix DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
EXECUTIVE SUMMARY
Principal Authors
Dr. William E. Wilson—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly Comfort—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. J.H.B. Garner—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Allan Marcus—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. James A. Raub—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
CHAPTER 1. INTRODUCTION
Principal Authors
Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
April 2002 I-xxx DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
CHAPTER 2. PHYSICS, CHEMISTRY, AND MEASUREMENT
OF PARTICULA TE MA TTER
Principal Authors
Dr. William Wilson—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Candis S. Claiborn—Washington State University, Laboratory for Atmospheric Research,
Department of Civil and Environmental Engineering, P.O. Box 642910, Pullman, WA 99164
Contributing Authors
Dr. Barbara J. Turpin, The State University of New Jersey, Rutgers, Department of
Environmental Sciences and Rutgers Cooperative Extension, New Brunswick, NJ 08901-8551
Dr. James J. Schauer, University of Wisconsin, College of Engineering, Department of Civil and
Environmental Engineering, Madison, WI 53706
Dr. Judith C. Chow, Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512
Dr. John G. Watson, Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512
Dr. Tom Cahill, University of California, Davis, One Shields Ave., Davis, CA 95616
Dr. Timothy Buckley—Johns Hopkins University, Department of Environmental Health
Sciences, 615 North Wolfe Street,, Baltimore, MD 21205
Ms. Lee Byrd—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C,
Petaluma, CA 94549
Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430,
Irvine, C A 92612
Dr. Delbert Eatough—Brigham Young University, E 114 BNSN,
Department of Chemistry and Biochemistry, Provo, UT 84602
April 2002 I-xxxi DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Edward O. Edney—National Exposure Research Laboratory (MD-84)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Sheldon Friedlander—University of California at Los Angeles, Department of Chemical
Engineering, 5531 Boelter Hall, Los Angeles, CA 90095
Dr. Judith Graham—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lynn Hildemann—Stanford University, Civil and Environmental Engineering Department,
Stanford, CA 94305
Mr. Jim Homolya—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Rudolf Husar—CAPITA, Washington University, Campus Box 1124,
One Brookings Drive, St. Louis, MO 63130
Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Scott Mathias—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom McCurdy—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Frank McElroy—National Exposure Research Laboratory (MD-46),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Haluk Ozkaynak—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom Pace—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
April 2002 I-xxxii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Joseph Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Richard Poirot—VT Air Program, Building 3 South, 103 South Main Street,
Waterbury, VT 05671
Dr. Linda Sheldon—National Exposure Research Laboratory (MD-77),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Helen Suh—Harvard School of Public Health, 665 Huntington Avenue,
Boston, MA 02461
Mr. Robert Wayland—Office of Air Quality Planning and Standards (MD-15),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Russell Weiner—National Exposure Research Laboratory (MD-46)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street,
Sacramento, CA 95814
CHAPTER 3. CONCENTRATIONS, SOURCES, AND
EMISSIONS OF A TMOSPHERIC PARTICVLA TE MA TTER
Principal Author
Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributing Authors
Dr. Barbara J. Turpin, The State University of New Jersey, Rutgers, Department of
Environmental Sciences and Rutgers Cooperative Extension, New Brunswick, NJ 08901-8551
Dr. James J. Schauer, University of Wisconsin, College of Engineering, Department of Civil and
Environmental Engineering, Madison, WI 53706
April 2002 I-xxxiii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
Dr. JoEllen Bandemeyer—Research Triangle Institute, PO Box 12194, Research Triangle
Park,NC 27709
Dr. Timothy Buckley—Johns Hopkins University, Department of Environmental Health
Sciences, 615 North Wolfe Street, Baltimore, MD 21205
Ms. Lee Byrd—Office of Air Quality Planning and Standards (MD-14),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C,
Petaluma, CA 94549
Dr. Candis S. Claiborn—Washington State University, Laboratory for Atmospheric Research,
Department of Civil and Environmental Engineering, P.O. Box 642910, Pullman, WA 99164
Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430,
Irvine, C A 92612
Mr. Tom Coulter—National Exposure Research Laboratory (MD-47),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Delbert Eatough—Brigham Young University, E 114 BNSN,
Department of Chemistry and Biochemistry, Provo, UT 84602
Dr. Edward O. Edney—National Exposure Research Laboratory (MD-47), U.S. Environmental
Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Terence Fitz-Simons—Office of Air Quality Planning and Standards (MD-14),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Sheldon Friedlander—University of California at Los Angeles, Department of Chemical
Engineering, 5531 Boelter Hall, Los Angeles, CA 90095
Dr. Judith Graham—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lynn Hildemann—Civil and Environmental Engineering Department,
Stanford University, Stanford, CA 94305
April 2002 I-xxxiv DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Mr. Jim Homolya—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Rudolf Husar—CAPITA, Washington University, Campus Box 1124,
One Brookings Drive, St. Louis, MO 63130
Dr. Allen S. Lefohn—A.S.L. & Associates, 111 North Last Chance Gulch, Helena, MT 59601
Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Phil Lorang—Office of Air Quality Planning and Standards (MD-14),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Karen Magliano—California Air Resources Board, 2020 L Street,
Sacramento, CA 95814
Mr. Scott Mathias—Office of Air Quality Planning and Standards (MD-15)
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom McCurdy—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Frank McElroy—National Exposure Research Laboratory (MD-46),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Nehzat Mottallebi—California Air Resources Board, 2020 L Street,
Sacramento, CA 95814
Dr. Haluk Ozkaynak—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom Pace—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Richard Poirot—VT Air Program, Building 3 South, 103 South Main Street,
Waterbury, VT 05671
April 2002 I-xxxv DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Mr. Win Setiawan—California Air Resources Board, 2020 L Street,
Sacramento, CA 95814
Dr. Linda Sheldon—National Exposure Research Laboratory (MD-77),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Helen Suh—Harvard School of Public Health, 665 Huntington Avenue,
Boston, MA 02461
Mr. Robert Wayland—Office of Air Quality Planning and Standards (MD-15),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street, Sacramento, CA 95814
Dr. William Wilson—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
CHAPTER 4. ENVIRONMENTAL EFFECTS OF PARTICULA TE MA TIER
Principal Authors
Ms. Beverly Comfort—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. J.H.B. Garner—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David A. Grantz—University of California/Riverside, Kearney Agricultural Center,
9240 South Riverbend Avenue, Parlier, CA 93648
April 2002 I-xxxvi DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Principal Authors
(cont'd)
Dr. Paul J. Hanson—Environmental Sciences Division, Oak Ridge National Laboratory,
P.O. Box 2008, Bethel Valley Road, Building 1059, Oak Ridge, TN 37831-6422
Dr. Dale W. Johnson—Environmental and Resource Science, 1000 Valley Road, University of
Nevada, Reno, NV 89512
Dr. Joseph P. Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William H. Smith—Yale University School of Forestry and Environmental Studies,
370 Prospect Street, New Haven, CT 06511
Contributors and Reviewers
Dr. Larry T. Cupitt—National Exposure Research Laboratory (MD-75),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Debra Meyer Wefering—Duckterather Weg 61, Bergisch Gladbach, Germany 54169
(formerly with National Exposure Research Laboratory [MD-56], U.S. Environmental Protection
Agency, Research Triangle Park, NC 27711)
Dr. Russell R. Dickerson—University of Maryland, Department of Meteorology,
Stadium Drive, College Park, MD 20742
Dr. Anne Grambsch—National Center for Environmental Assessment (860ID),
U. S. Environmental Protection Agency, Washington, DC 20036
Dr. Sagar V. Krupa—University of Minnesota, Department of Plant Pathology,
495 BorlaugHall, 1991 Upper Buford Circle, St. Paul, MN 55108
Dr. Alan J. Krupnick—Quality of the Environment Division, Resources for the Future,
1616 P Street, NW, Washington, DC 20036
Mr. Paul T. Roberts—Sonoma Technology, Inc.,1360 Redwood Way - Suite C,
Petaluma, CA 94954
Mr. John Spence—1206 Sturdivant Drive, Cary, NC 27511
April 2002 I-xxxvii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Richard Zepp—National Exposure Research Laboratory (IOD),
U. S. Environmental Protection Agency, Athens, GA
CHAPTER 5. HUMAN EXPOSURE TO PARTICULA TE MA TIER
AND ITS CONSTITUENTS
Principal Authors
Dr. David T. Mage—Institute for Survey Research, Temple University,
Philadelphia, PA 19122-6099 (formerly with the National Exposure Research
Laboratory (MD-56), U.S. Environmental Protection Agency, Research Triangle
Park,NC 27711)
Mr. Thomas McCurdy—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Linda S. Sheldon—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Haluk Ozkaynak—National Exposure Research Laboratory (MD-56),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. William E. Wilson—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributing Authors
Dr. Janet Burke—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Roy Fortmann—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Gary Norris—National Exposure Research Laboratory (MD-47),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
April 2002 I-xxxviii DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributing Authors
(cont'd)
Dr. Anne Rea—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Alan Vette—National Exposure Research Laboratory (MD-56),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Contributors and Reviewers
Dr. Timothy Buckley—Johns Hopkins University, Department of Environmental Health
Sciences, 615 North Wolfe Street,, Baltimore, MD 21205
Ms. Lee Byrd—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C, Petaluma, CA 94549
Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430,
Irvine, C A 92612
Dr. Delbert Eatough—Brigham Young University, E 114 BNSN,
Department of Chemistry and Biochemistry, Provo, UT 84602
Dr. Sheldon Friedlander—University of California at Los Angeles, Department of Chemical
Engineering, 5531 Boelter Hall, Los Angeles, CA 90095
Dr. Judith Graham—National Exposure Research Laboratory (MD-75),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Lester D. Grant—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Vic Hasselblad—29 Autumn Woods Drive, Durham, NC 27713
Dr. Lynn Hildemann—Civil and Environmental Engineering Department
Stanford University, Stanford, CA 94305
Mr. Jim Homolya—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
April 2002 I-xxxix DRAFT-DO NOT QUOTE OR CITE
-------
Authors, Contributors, and Reviewers
(cont'd)
Contributors and Reviewers
(cont'd)
Dr. Rudolf Husar—CAPITA, Washington University, Campus Box 1124,
One Brookings Drive, St. Louis, MO 63130
Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Charles W. Lewis—National Exposure Research Laboratory (MD-47),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Allan Marcus—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Frank McElroy—National Exposure Research Laboratory (MD-46),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Tom Pace—Office of Air Quality Planning and Standards (MD-14),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Joseph Pinto—National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Richard Poirot—VT Air Program, Building 3 South, 103 South Main Street,
Waterbury, VT 05671
Mr. Harvey Richmond—Office of Air Quality Planning and Standards (MD-15),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Linda Sheldon—National Exposure Research Laboratory (MD-77),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. Helen Suh—Harvard School of Public Health, 665 Huntington Avenue,
Boston, MA 02461
Mr. Robert Wayland—Office of Air Quality Planning and Standards (MD-15),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street, Sacramento, CA 95814
Dr. Jim Xue—Harvard School of Public Health, 665 Huntington Avenue, Boston, MA 02115
April 2002 I-xl DRAFT-DO NOT QUOTE OR CITE
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
FOR PARTICULATE MATTER
Executive Direction
Dr. Lester D. Grant—Director, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Scientific Staff
Dr. William E. Wilson—PM Team Leader and Air Quality Coordinator, Physical Scientist,
National Center for Environmental Assessment (MD-52), U.S. Environmental Protection
Agency, Research Triangle Park, NC 27711
Dr. Lawrence J. Folinsbee—Health Coordinator, Chief, Environmental Media Assessment
Group, National Center for Environmental Assessment (MD-52), U.S. Environmental Protection
Agency, Research Triangle Park, NC 27711 (now deceased)
Dr. Dennis J. Kotchmar—Project Manager, Medical Officer, National Center for Environmental
Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Dr. Robert Chapman—Medical Officer, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Beverly Comfort—Health Scientist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. William Ewald—Health Scientist, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. J.H.B. Garner—Ecological Scientist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. David Mage—Physical Scientist, Institute for Survey Research, Temple University,
Philadelphia, PA 19122-6099
Dr. Allan Marcus—Statistician, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Dr. James McGrath—Visiting Senior Health Scientist, National Center for Environmental
Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
April 2002 I-xli DRAFT-DO NOT QUOTE OR CITE
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
FOR PARTICULATE MATTER
(cont'd)
Scientific Staff
(cont'd)
Dr. Joseph P. Pinto—Physical Scientist, National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. James A. Raub—Health Scientist, National Center for Environmental Assessment (MD-52),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711
Technical Support Staff
Mr. Randy Brady—Deputy Directory, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Douglas B. Fennell—Technical Information Specialist, National Center for Environmental
Assessment (MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711
Ms. Emily R. Lee—Management Analyst, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Diane H. Ray—Program Specialist, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Ms. Eleanor Speh—Office Manager, Environmental Media Assessment Branch, National Center
for Environmental Assessment (MD-52), U.S. Environmental Protection Agency, Research
Triangle Park, NC 27711
Ms. Donna Wicker—Administrative Officer, National Center for Environmental Assessment
(MD-52), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
Mr. Richard Wilson—Clerk, National Center for Environmental Assessment (MD-52),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
April 2002 I-xlii DRAFT-DO NOT QUOTE OR CITE
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
FOR PARTICULATE MATTER
(cont'd)
Document Production Staff
Dr. Carol A. Seagle—Technical Editor, Computer Sciences Corporation,
2803 Slater Road, Suite 220, Morrisville, NC 27560
Ms. Diane G. Caudill—Graphic Artist, Computer Sciences Corporation,
2803 Slater Road, Suite 220, Morrisville, NC 27560
Ms. Carolyn T. Perry—Word Processor, Computer Sciences Corporation,
2803 Slater Road, Suite 220, Morrisville, NC 27560
Ms. Kelly Quifiones—Word Processor, InfoPro, Inc., 8405 Colesville Road, 2nd Floor,
Silver Spring, MD 20910
Technical Reference Staff
Mr. John A. Bennett—Technical Information Specialist, SANAD Support Technologies, Inc.,
11820 Parklawn Drive, Suite 400, Rockville, MD 20852
Ms. Sandra L. Hughey—Technical Information Specialist, SANAD Support Technologies, Inc.,
11820 Parklawn Drive, Suite 400, Rockville, MD 20852
Ms. Beth Olen—Records Management Technician, Reference Retrieval and Database Entry
Clerk, InfoPro, Inc., 8405 Colesville Road, 2nd Floor, Silver Spring, MD 20910
April 2002 I-xliii DRAFT-DO NOT QUOTE OR CITE
-------
Abbreviations and Acronyms
oabs light-absorption coefficient
oag light-absorption coefficient of gases
oap light-absorption coefficient of particles
°~ext
light-extinction coefficient
og geometric standard deviation
"scat
light-scattering coefficient
osg light-scattering coefficient of gases
osp light-scattering coefficient of particles
4-POBN
A
AAS
ACGffl
a-(4-pyridyl- 1 -oxide)-N-tert-butylnitrone
alveolar
atomic absorption spectrophotometry
American Conference of Governmental Industrial Hygienists
AD
ADS
AES
AIRS
AM
AQCD
AQI
AQRV
ARIES
ASOS
ATOM
ATOFMS
annular denuder system
atomic emission spectroscopy
Aerometric Information Retrieval System
alveolar macrophages
Air Quality Criteria Document
Air Quality Index
Air Quality Related Values
Aerosol Research and Inhalation Epidemiology Study
Automated Surface Observing System
aerosol and toxic deposition model
time-of-flight mass spectrometer
b
Ba
absorption coefficient
April 2002
I-xliv
DRAFT-DO NOT QUOTE OR CITE
-------
BAD
brachial artery diameter
BAL
bronchoalveolar lavage
BALF
bronchoalveolar lavage fluid
BAUS
brachial artery ultrasonography
BC
black carbon (see also CB)
BW
bronchial wash
BYU
Bringham Young University
C
apparent contrast
Ca+
calcium
CAA
Clean Air Act
CAAM
continuous ambient mass monitor
CAMNET
CAPs
concentrated ambient particles
CARS
California Air Resources Board
CASAC
Clean Air Scientific Advisory Committee
CASTNet
Clean Air Status and Trends Network
CAT
computer-aided tomography
CB
carbon black
base cation
CC
carbonate carbon
CC14
carbon tetrachloride
CCPM
continuous coarse particle monitor
CCSEM
computer-controlled scanning electron microscopy
CEN
European Standardization Committee
CF
Cystic Fibrosis
CFA
coal fly ash
CFCs
chlorofluorocarbons
CFD
computational fluid dynamics
April 2002
I-xlv
DRAFT-DO NOT QUOTE OR CITE
-------
CFR
CH2O
GIF
CL
CMAQ
CMB
CMD
CMP
CMSA
C0
CO
CO CD
COPD
CPC
CPZ
CR
CRP
Code of Federal Regulations
formaldehyde
charcoal-impregnated cellulose fiber
chemiluminescence
Community Multi-Scale Air Quality
chemical mass balance
count mean diameter
copper smelter dust
Consolidated Metropolitan Statistical Area
initial contrast
carbon monoxide
Air Quality Criteria Document for Carbon Monoxide
chronic obstructive pulmonary disease
condensation particle counter
capsazepine
concentration-response
Coordinated Research Program
CSIRO
CSMCS
CTM
CV
Carbonaceous Species Methods Comparison Study
chemistry-transport model
coefficient of variation
D5o
Da
DAQM
DCFH
DE
DE
DEF
Denver Air Quality Model
dichlorofluorescin
deposition efficiencies
diesel exhaust
Deferoxamine
April 2002
I-xlvi
DRAFT-DO NOT QUOTE OR CITE
-------
DEP
diesel exhaust particles
DHR
dihydrorhodamine-123
DMS
dimethyl sulfide
DMTU
dimethylthiourea
DOFA
domestic oil fly ash
DPM
diesel particulate matter
DRG
dorsal root ganglia
dv
deciview index
BAD
electrical aerosol detector
EC
elemental carbon
ECAO
Environmental Criteria and Assessment Office
ECG
electrocardiogram
EDXRF
energy dispersive X-ray fluorescence
EGA
evolved gas analysis
EGF
epidermal growth factor
ELSIE
Elastic Light Scattering and Interactive Efficiency
ERK
extracellular receptor kinase
ESP
electrostatic precipitator
ESR
electron spin resonance
ET
extrathoacic
ETS
environmental tobacco smoke
EU
endotoxin units
EXPOLIS
flux
FEF
forced expiratory flow
FEVj
forced expiratory volume in 1 second
FID
flame ionization detection
FMD
flow-mediated dilation
April 2002
I-xlvii
DRAFT-DO NOT QUOTE OR CITE
-------
FPD
flame photometric detector
FRM
Federal Reference Method
gS02
gaseous sulfur dioxide
GC
gas chromatography
GCMs
General Circulation Models
GCVTC
Grand Canyon Visibility Transport Commission
GG/MSD
gas chromatography/mass-selective detection
GHG
greenhouse gases
GMCSF
granulocyte macrophage colony stimulating factor
GMPD
geometric mean particle diameter
GSD
geometric standard deviation (see also o )
GSH
glutathione
H2SO4
sulfuric acid
HAAQS
HDM
house dust mite
HDS
honeycomb denuder/filter pack sampler
HEADS
Harvard-EPA Annular Denuder Sampler
HEI
Health Effects Institute
hivol
High blume sampler
HNO3
nitric acid
HR
heart rate
HTGC-MS
high temperature gas chromotography-mass spectrometry
radiance
inhibitory kappa B alpha
apparent radiance of the background
transmitted radiance of the background
1C
ion chromatography
ICAM-1
intercellular adhesion molecule-1
April 2002
I-xlviii
DRAFT-DO NOT QUOTE OR CITE
-------
ICP
inductively coupled plasma
ICRP
International Commission on Radiological Protection
le
equilibrium radiance or source function
IPS
Integrated Forest Study
IgE
immunoglobin E
IgG
immunoglobin G
IL
interleukin
IMPROVE
Interagency Monitoring of Protected Visual Environments
INAA
instrumental neutron activation analysis
IOVPS
integrated organic vapor/particle sampler
intraperitoneal
path radiance
IPCC
Intergovernmental Panel on Climate Change
IPM
inhalable paniculate matter
IPN
Inhalable Paniculate Network
ISO
International Standards Organization
transmitted radiance
INK
c-jun N-terminal kinase
'scp
light scattering by coarse particles
Jsfp
light scattering by fine particles
Jspd
light scattering coefficient of particles under dry conditions
'spw
light scattering coefficient of particles under humid conditions
K
Koschmieder constant
K+
potassium ion
KOH
potassium hydroxide
LAI
leaf area indices
LFA-1
leukocyte function-associated antigen-1
LN
lymph nodes
April 2002
I-xlix
DRAFT-DO NOT QUOTE OR CITE
-------
LoS
1pm, Lpm, L/min
LPS
LWCA
MAA
MAACS
MADPro
MAPK
MAQSIP
MCM
MCT
MEK
MIP
Mm
MMAD
MMD
MMPs
MOUDI
MPL
MPO
MS
MSA
MSAs
MSH
low sulfur
liters per minute
lipopolysaccharide
liquid water content analyzer
mineral acid anion
Metropolitan Acid Aerosol Characterization Study
Mountain Acid Deposition Program
mitogen-activated protein kinase
page 3-83
mass concentrations monitor
monocrotaline
mitogen-activated protein kinase
macrophage inflammatory protein
megameters
mean median aerodynamic diameter (see og)
mass median diameter
matrix metalloproteinases
micro-orifice uniform deposit impactor
multipath lung
myeloperoxidase
mass spectroscopy
methane sulfonic acid
metropolitan statistical areas
Mount St. Helens
MSP
NAC
NAL
NAMS
N-acetylcysteine (antioxidant)
nasal lavage fluid
National Ambient Monitoring Stations
April 2002
1-1
DRAFT-DO NOT QUOTE OR CITE
-------
NaN3
sodium azide
NAPAP
National Acid Precipitation Assessment Program
NAPRMN
NARSTO
NAST
National Assessment Synthesis Team
NCRPM
National Council on Radiation Protection and Measurements
ND
NIST diesel (also, not determined)
NDDN
National Dry Deposition Network
NDIR
nondispersive infrared spectrophotometry
NESCAUM
Northeast States for Coordinated Air Use Management
NF
nuclear factor
NF-KB
nuclear factor kappa B
NFRAQS
North Frontal Range Air Quality Study
NH3
ammonia
NIL
ammonium
(NH4)2 S04
ammonium sulfate
NILH,S(X
ammonium acid sulfate
NHBE
normal human bronchial epithelial
NIOSH
NIR
NIST
National Institute of Standards and Technology
NMD
nitroglycerine-mediated dilation
NMD
number mean diameter
NMRI
Naval Medical Research Institute
NO
nitrogen oxide
NO,
nitrogen dioxide
NO3-
nitrate
NOPL
naso-oro-pharyngo-laryngeal
April 2002
I-li
DRAFT-DO NOT QUOTE OR CITE
-------
NOX
NPP
NRC
NuCM
03
OAA
OAQPS
OAR
OC
OFA
CH-
ORD
OVA
P
P so42-
PAH
PAHs
PAN
PAR
PB
PEL
nitrogen oxides
net primary production
National Research Council
nutrient cycling model
ozone
Ottowa ambient air
Office of Air Quality Planning and Standards
Office of Air and Radiation
organic carbon
oil fly ask
hydroxyl ion
Office of Research and Development
ovalbumin
partial pressure
particulate sulfate
polynuclear aromatic hydrocarbon
polycyclic aromatic hydrocarbons
peroxyacetyl nitrate
photosynthetically active radiation
polymyxin-B
planetary boundary layer
PBY
PC
PC
PC-BOSS
PCA
PCBs
PCDD
April 2002
pyrolitic carbon
particle concentrator
Particulate Concentrator-Brigham Young University Organic
Sampling System
principal component analysis
polychloronated biphenyls
polychlorinated dibenzo-p-dioxins
I-lii DRAFT-DO NOT QUOTE OR CITE
-------
PCDF
polychlorinated dibenzofurans
PCM
particle composition monitor
pdf
probability density functions
PDGF
platelet-derived growth factor
PEM
Personal Environmental Monitor
PESA
proton elastic scattering analysis
PFA
PIXE
proton induced X-ray emission
PM
particulate matter
PM AQCD
PM Air Quality Criteria Document
PM(10.25)
coarse particulate matter
PM
-2.5
fine particulate matter
PMF
positive matrix factorization
PMN
polymorphonuclear leukocytes
equilibrium vapor pressure
poly I:C
polyionosinic-polycytidilic acid
POP
persistent organic pollutant
PROBDET
Probability of Detection Algorithm
PTEAMS
PTEP
PM10 Technical Enhancement Program
PTFE
polytetrafluoroethylene
PTFE
polytetrafluoroethylene
PUF
polyurethane foam
Q
respiratory flow rates
Qabs
efficiency of absorption
Qext
efficiency of extinction
Qscat
efficiency of scattering
aerodynamic resistance
April 2002
I-liii
DRAFT-DO NOT QUOTE OR CITE
-------
RAAS
RADM
Regional Acid Deposition Model
RAMS
Real-Time Air Monitoring System
RAMS
Regional Air Monitoring Study
RAPS
Regional Air Pollution Study
boundary layer resistance
canopy resistance
REMSAD
Regulatory Modeling System for Aerosols and Deposition
RFC
residual fuels oils
RH
relative humidity
ROFA
residual oil fly ash
ROFA
residual oil fly ash
ROME
Reactive and Optics Model Emissions
ROS
reactive oxygen species
RPM
respirable particulate matter
RPM
Regional Particulate Model
RTE
rat tracheal epithelial
RTF
Research Triangle Park
SASS
sec
saturation ratio
SA
Sierra Anderson
SAD
small airway disease
SCAQS
Southern California Air Quality Study
scos
Southern California Ozone Study
sd
standard deviation
SEM
scanning electron microscopy
SES
sample equilibration system
April 2002
I-liv
DRAFT-DO NOT QUOTE OR CITE
-------
SEV
Sensor Equivalent Visibility
SH
spontaneously hypertensive
SIP
State Implementation Plans
SIXE
synchrotron induced X-ray emission
SL
stochastic lung
SLAMS/NAMS
SLAMS
State and Local Air Monitoring Stations
SLE
St. Louis encephalitis
SMPS
scanning mobility particle sizer
SO,
sulfur dioxide
scx2-
sulfate
SOA
SOC
semivolatile organic compounds
SoCAB
South Coast Air Basin
SOD
superoxide dismutase
SOPM
secondary organic particulate matter
SP
Staff Paper
SPM
synthetic polymer monomers
SRI
SRM
standard reference method
SSM
solid sampler module
Stk
Stokes number
SUVB
solar ultraviolet B radiation
svoc
semivolatile organic compounds
SWMMC
Southwest Metropolitan Mexico City
T(CO)
core temperature
TB
tracheabronchial
TDF
total deposition fraction
April 2002
I-lv
DRAFT-DO NOT QUOTE OR CITE
-------
TDMA
Tandem Differential Mobility Analyzer
TEOM
tapered element oscillating microbalance
TEOMs
TIMP
tissue inhibitor of metaloproteinase
TLN
TNF
tumor necrosis factor
TOFMS
aerosol time-of-flight mass spectroscopy
TOR
thermal/optical reflectance
TOT
thermal/optical transmission
TPM
thoracic paniculate matter
TRXRF
total reflection X-ray fluorescence
TSI
TSP
total suspended particulates
UAM-V
Urban Airshed Model Version V
UCM
unresolved complex mixture
ufCB
ultrafine carbon black
UFP
ultrafine fluorospheres
UNEP
United Nations Environment Programme
URG
University Research Glassware
USGCRP
U.S. Global Change Research Program
UVD
Utah Valley dust
VAPS
Versatile Air Pollution Samplers
VASM
Visibility Assessment Scoping Model
VBE
Japanese B encephalitis
VCAM-1
vascular cell adhesion molecule-1
deposition velocity
VDI
VOC
volatile organic compounds
April 2002
I-lvi
DRAFT-DO NOT QUOTE OR CITE
-------
vs
v,
v,
we
WEE
WINS
WIS
WKY
WMO
Wo
WRAC
X-XRF
XAD
XRF
sedimentation velocity
turbulent diffusion velocity
tidal volume
tungsten carbide
western equine encephalitis
Well Impactor Ninety- Six
Wistar
Wi star-Kyoto
World Meteorological Organization
single scattering albedo
Wide Range Aerosol Classifier
synchrotron induced X-ray fluorescence
polystyrene-divinyl benzene
X-ray fluorescence
V*
April 2002
I-lvii
DRAFT-DO NOT QUOTE OR CITE
-------
i EXECUTIVE SUMMARY
2
3
4 E.I INTRODUCTION
5 E. 1.1 Purpose of the Document
6 The purpose of this document, Air Quality Criteria for Particulate Matter, is to present air
7 quality criteria for parti culate matter (PM) in accordance with Clean Air Act (CAA) Sections 108
8 and 109, which govern establishment, review, and revision of U.S. National Ambient Air Quality
9 Standards (NAAQS) as follows:
10
11 • Section 108 directs the U.S. Environmental Protection Agency (EPA) Administrator to list
12 pollutants that may reasonably be anticipated to endanger public health or welfare and to issue
13 air quality criteria for them. The air quality criteria are to reflect the latest scientific
14 information useful in indicating the kind and extent of all identifiable effects on public health
15 and welfare expected from the presence of the pollutant in ambient air.
16
17 • Section 109 directs the EPA Administrator to set and periodically revise, as appropriate,
18 (a) primary NAAQS, which in the judgement of the Administrator, are requisite to protect
19 public health, with an adequate margin of safety, and (b) secondary NAAQS which, in the
20 judgement of the Administrator, are requisite to protect the public welfare from any known or
21 anticipated adverse effects (e.g., impacts on vegetation, crops, ecosystems, visibility, climate,
22 man-made materials, etc.).
23
24 • Section 109 of the CAA also requires periodic review and, if appropriate, revision of existing
25 criteria and standards. Also, an independent committee of non-EPA experts, the Clean Air
26 Scientific Advisory Committee (CASAC), is to provide the EPA Administrator advice and
27 recommendations regarding the scientific soundness and appropriateness of criteria and
28 NAAQS.
29
April 2002 E-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 To meet these CAA mandates, this document assesses the latest scientific information
2 useful in deriving criteria as scientific bases for decisions on possible revision of current
3 PM NAAQS. A separate EPA PM Staff Paper will draw upon assessments in this document,
4 together with technical analyses and other information, to identify alternatives for consideration
5 by the EPA Administrator with regard to possible retention or revision of the PM NAAQS.
6
7 E. 1.2 Organization of the Document
8 The present document is organized into nine chapters, as follows:
9
10 • This Executive Summary summarizes key points from the ensuing chapters.
11
12 • Chapter 1 provides a general introduction, including a brief summary of the history of the PM
13 NAAQS and an overview of issues, methods and procedures used to prepare this document.
14
15 • Chapters 2, 3 and 5 provide background information on air quality and exposure aspects to help
16 to place the succeeding discussions of PM effects into perspective.
17
18 • Chapter 4 deals with environmental effects of PM on vegetation and ecosystems, visibility,
19 manmade materials, and climate.
20
21 • Human health issues related to PM are addressed in Chapter 6 (on dosimetry); Chapter 7 (on
22 toxicology); and Chapter 8 (on community epidemiology).
23
24 • Chapter 9 provides an integrative synthesis of key points from the preceding chapters.
25
26
27 E.2 AIR QUALITY AND EXPOSURE ASPECTS
28 The document's discussion of air quality and exposure aspects considers chemistry and
29 physics of atmospheric PM; analytical techniques for measuring PM mass, size, and chemical
30 composition; sources of ambient PM in the United States; temporal/spatial variability and trends
April 2002 E-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 in ambient U.S. PM levels; and ambient concentration-human exposure relationships. Key
2 findings are summarized in the next several sections. Overall, the atmospheric science and air
3 quality information provides further evidence substantiating the 1996 PM AQCD conclusion that
4 distinctions between fine and coarse mode particles (in terms of emission sources, formation
5 mechanisms, atmospheric transformation, transport distances, air quality patterns, and exposure
6 relationships) warrant fine and coarse PM being viewed as separate subclasses of ambient PM.
7
8 E.2.1 Chemistry and Physics of Atmospheric Particles
9 • Airborne PM is not a single pollutant, but rather is a mixture of many subclasses of pollutants
10 with each subclass containing many different chemical species. Atmospheric PM occurs
11 naturally as fine-mode and coarse-mode particles that, in addition to falling into different size
12 ranges, differ in formation mechanisms, chemical composition, sources, and exposure
13 relationships.
14
15 • PM may be primary or secondary. PM is called primary if it exists in the same chemical form
16 in which it was emitted or generated. PM is called secondary if it is formed through the
17 atmospheric reaction of a precursor gas that forms a condensible product that in turn nucleates
18 to form new particles or condenses on existing particles.
19
20 • Fine-mode PM is derived primarily from combustion material that has volatilized and then
21 condensed to form primary PM or from precursor gases reacting in the atmosphere to form
22 secondary PM. New fine-mode particles are formed by the nucleation of gas phase species;
23 they grow by coagulation (existing particles combining) or condensation (gases condensing on
24 existing particles). Fine particles are composed of freshly generated nuclei-mode particles, also
25 called ultrafme or nanoparticles, and an accumulation mode (so-called because particles grow
26 into and remain in that mode).
27
28 • Coarse-mode PM, in contrast, is formed by crushing, grinding, and abrasion of surfaces, which
29 breaks large pieces of material into smaller pieces. These particles are then suspended by the
30 wind or by anthropogenic activity. Energy considerations limit the break-up of large mineral
31 particles and small particle aggregates generally to a minimum size of about 1 //m in diameter,
April 2002 E-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 although biological material may exist or fragment into smaller sizes. Mining and agricultural
2 activities are examples of anthropogenic sources of coarse-mode particles. Fungal spores,
3 pollen, and plant and insect fragments are examples of natural bioaerosols also suspended as
4 coarse-mode particles.
5
6 • Within atmospheric particle modes, the distribution of particle number, surface, volume, and
7 mass by diameter is frequently approximated by lognormal distributions. Aerodynamic
8 diameter, which depends on particle density and is defined as the diameter of a particle with the
9 same settling velocity as a spherical particle with unit density (1 g/cm3), is often used to
10 describe particle size. Typical values of the mass median aerodynamic diameters (MMAD) are
11 0.05 to 0.07 //m for the nuclei mode, 0.3 to 0.7 //m for the accumulation mode, and 6 to 20 //m
12 for the coarse mode. At high relative humidities or in air containing evaporating fog or cloud
13 droplets, the accumulation mode may be split into a droplet mode (MMAD = 0.5 to 0.8 //m)
14 and a condensation mode (MMAD = 0.2 to 0.3 //m).
15
16 • Research studies use impactors to determine mass and composition as a function of size over a
17 wide range and particle counting devices to determine number of particles as a function of size.
18 Such studies indicate an atmospheric bimodal distribution of fine and coarse particle mass with
19 a minimum in the distribution between 1 and 3 //m aerodynamic diameter. Routine monitoring
20 studies prior to 1999 generally measured thoracic PM, i.e., PM10 (upper size limited by a 50%
21 cut at 10 //m aerodynamic diameter). Research studies and monitoring studies since 1999
22 measure fine PM, i.e., PM25 (upper size limited by a 50% cut point at 2.5 //m aerodynamic
23 diameter) and coarse thoracic PM, i.e, PM10_25 the coarse fraction of PM10, measured as the
24 difference between PM10 and PM2 5 mass measurements obtained at the same time and location
25 and with similar inlets and other sampling and handling specifications. Cut points are not
26 perfectly sharp for any of these PM indicators; some particles larger than the 50% cutpoint are
27 collected and some particles smaller than the 50% cutpoint are not retained.
28
29 • The terms "fine" and "coarse" were originally intended to apply to the two major atmospheric
30 particle distributions which overlap in the size range between 1 and 3 //m diameter. Now, fine
31 has come to be often associated with the PM2 5 fraction and coarse is often used to refer to
April 2002 E-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 PM10_2.5, coarse thoracic PM. However, PM2 5 may also contain, in addition to the fine-particle
2 mode, some of the lower-size tail of the coarse particle mode between about 1 and 2.5 //m
3 aerodynamic diameter. Conversely, under high relative humidity conditions, the larger fine
4 particles in the accumulation mode may also extend into the 1 to 3 //m aerodynamic diameter
5 range.
6
7 • Four approaches are used to classify particles by size: (1) modes, based on formation
8 mechanisms and the modal structure observed in the atmosphere, e.g., nuclei and accumulation
9 modes (which comprise the fine-particle mode) and the coarse-particle mode; (2) cut point,
10 based on the 50% cut point of the specific sampling device; (3) dosimetry, based on the ability
11 of particles to enter certain regions of the respiratory tract; and (4) regulatory, based on
12 instrument configuration or 50% cut-points, e.g., high volume sampler, PM10, and PM2 5.
13
14 E.2.2 Sources of Airborne Particles in the United States
15 • The chemical complexity of airborne particles requires that the composition and sources of a
16 large number of primary and secondary components be considered. Major components of fine
17 particles are: sulfate, strong acid, ammonium, nitrate, organic compounds, trace elements
18 (including metals), elemental carbon, and water.
19
20 • Primary particles are emitted directly from sources. Secondary particles are formed from
21 atmospheric reactions of sulfur dioxide (SO2), nitrogen oxides (NOX), and certain organic
22 compounds. NO reacts with ozone (O3) to form NO2. SO2 and NO2 react with hydroxy radical
23 (OH) during the daytime to form sulfuric and nitric acid. During the nighttime, NO2 reacts
24 with ozone and forms nitric acid through a sequence of reactions involving the nitrate radical
25 (NO3). These acids may react further with ammonia to form ammonium sulfates and nitrates.
26 Some types of higher molecular weight organic compounds react with OH radicals, and olefinic
27 compounds also react with ozone to form oxygenated organic compounds, which nucleate or
28 can condense onto existing particles. SO2 also dissolves in cloud and fog droplets, where it
29 may react with dissolved O3, H2O2, or, if catalyzed by certain metals, with O2, yielding sulfuric
30 acid or sulfates, that lead to PM when the droplet evaporates.
31
April 2002 E-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Organic compounds constitute from 10 to 70% of dry PM25 mass. Whereas the chemistry of
2 particulate nitrate and sulfate formation has been relatively well studied, the chemistry of
3 secondary organic particulate matter formation is still under active investigation. Although
4 additional sources of secondary organic PM might still be identified, there appears to be a
5 general consensus that biogenic compounds (monoterpenes, sesquiterpenes) and aromatic
6 compounds (e.g., toluene and ethylbenzene) are the most significant precursors. Atmospheric
7 transformations of the compounds, which are formed in the particle phase during the aging of
8 particles, are still not adequately understood.
9
10 • Receptor modeling has proven to be a useful method for identifying contributions of different
11 types of sources, especially for the primary components of ambient PM. Apportionment of
12 secondary PM is more difficult because it requires consideration of atmospheric reaction
13 processes and rates. Results from western U.S. sites indicate that fugitive dust, motor vehicles,
14 and wood smoke are the major contributors to ambient PM samples there, whereas results from
15 eastern U.S. sites indicate that stationary combustion, motor vehicles and fugitive dust are
16 major contributors to ambient PM samples there. Sulfate and organic carbon are the major
17 secondary components in the East, while nitrates and organic carbon are the major secondary
18 components in the West.
19
20 E.2.3 Atmospheric Transport and Fate of Airborne Particles
21 • Primary and secondary fine particles have long lifetimes in the atmosphere (days to weeks) and
22 travel long distances (hundreds to thousands of kilometers). They tend to be uniformly
23 distributed over urban areas and larger regions, especially in the eastern United States. As a
24 result, they are not easily traced back to their individual sources.
25
26 • Coarse particles normally have shorter lifetimes (minutes to hours) and generally only travel
27 short distances (<10's of km). Therefore, coarse particles tend to be unevenly distributed
28 across urban areas and tend to have more localized effects than fine particles. However, dust
29 storms occasionally cause long range transport of small coarse-mode particles.
30
31
April 2002 E-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 E.2.4 Airborne Particle Measurement Methods
2 • Measurements of ambient PM mass and chemical composition are needed to determine
3 attainment of standards; to guide attainment of a standard (including determination of source
4 categories and validation of air quality models); and to determine health, ecological, and
5 radiative effects. A comprehensive approach requires a combination of analytical techniques to
6 assess: (1) mass, (2) crustal and trace elements, (3) water-soluble ionic species including
7 strong acidity, (4) elemental carbon, and (5) organic compounds.
8
9 • There are no calibration standards for suspended particle mass; therefore, the accuracy of
10 particle mass measurements cannot be definitively determined. The precision of particle mass
11 measurements can be determined by comparing results from collocated samplers. Mass
12 concentration measurements with a precision of 10% or better have been obtained with
13 collocated samplers of identical design. Field studies of EPA PM10 and PM25 reference
14 methods and reviews of field data from collocated PM10 and PM25 samplers show high
15 precision (better than ± 10%). The use of more careful techniques, including double weighing
16 of filters, can provide higher precision and may be needed for precise determination of PM10_25
17 by difference.
18
19 • Available technology allows accurate (± 10 to 15%) measurement of several of the major
20 components of coarse and fine particles (crustal and trace elements, sulfates, nitrates,
21 ammonium, and strong acidity). However, collection and measurement technologies for
22 elemental carbon and organic carbon are not as well established. The split between elemental
23 and organic carbon is operational, i.e., it is different for the two most frequently used
24 measurement techniques. In addition, in order to estimate the mass of organic PM, the ratio of
25 oxygen to carbon in organic PM must be estimated. It is higher for secondary organic than for
26 primary organic PM, adding further to the uncertainty in organic and elemental carbon
27 measurements.
28
29 • Semivolatile organic compounds and semivolatile ammonium compounds (such as NH4NO3)
30 may be lost by volatilization during sampling. Such losses may be very important in
31 woodsmoke impacted areas for organic compounds or in agricultural and other areas where low
April 2002 E-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 sulfate and high ammonia lead to high NH4NO3 concentrations. New techniques are now in
2 use for measurement of nitrates and new research techniques are being tested for measurement
3 of mass of semivolatile organic compounds in PM and of the total (semivolatile plus
4 non-volatile) PM mass. The Federal Reference Methods for PM10 and PM2 5 give precise
5 (± 10%) measurements of "equilibrated mass". However, the loss of semivolatile PM
6 (ammonium nitrate and organic compounds) and the possible retention of some particle-bound
7 water in current PM mass measurements contribute to uncertainly in the measurement of the
8 mass of PM as it exists suspended in the atmosphere.
9
10 • Intercomparisons, using different techniques and samplers of different designs, coupled with
11 mass balance studies (relating the sum of components to the measured mass), provide a method
12 for gaining confidence in the reliability of PM measurements.
13
14 E.2.5 Ambient PM Concentrations in the U.S.: Regional Patterns and Trends
15 • Particle mass data have been collected at a number of rural, suburban, and urban sites across
16 the United States by various local, state, and national programs. The data have been stored in
17 the Aerometric Information Retrieval System (AIRS). Data have also been collected at remote
18 sites as part of the IMPROVE and NESCAUM networks. An extensive analysis of this data
19 was reported in the 1996 Air Quality Criteria Document for Particulate Matter (PM AQCD).
20
21 • The median PM2 5 concentration across the United States during 1999 and 2000, the first two
22 years of operation of the PM25 FRM network, was 13 //g/m3, with a 95th percentile value of
23 18 //g/m3. The corresponding median PM10_25 concentration was 10 //g/m3, with a
24 95th percentile value of 21 Mg/m3.
25
26 • The spatial variability of PM2 5 concentrations is characterized in this document, based on the
27 availability of data at four or more sites within twenty-seven urban areas across the United
28 States. Correlations of PM25 concentrations between pairs of monitoring sites within the urban
29 areas examined ranges from low to high. Highest correlations are found at site-pairs that are
30 dominated by regional sources of secondary PM. Low correlations can be found if the sites are
31 located in different air sheds or if at least one of the sites is affected more strongly by local,
April 2002 E-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 primary sources. Although PM2 5 concentrations may be highly correlated between sites, the
2 concentrations themselves may not be uniform.
3
4 • Annual mean PM2 5 concentrations within the urban areas examined are typically within about
5 five Aig/m3 of each other. However, on a daily basis, absolute differences in PM25
6 concentrations can be much larger. In approximately half of these urban areas, the 90th
7 percentile difference in daily PM2 5 concentrations is greater than 10 //g/m3. Extreme values of
8 concentration differences were greater than 100 //g/m3 in a few cases. Caution should therefore
9 be exercised in using these data to approximate community-wide exposures.
10
11 • The database for characterizing the spatial variability of PM10_2 5 concentrations is not as
12 extensive as it is for PM2 5. Intersite correlations of PM10_2 5 concentrations were lower than
13 those for PM2 5 in the few urban areas that had sufficient data for both PM25 and PM10_2 5.
14 PM10_2 5 concentrations also tended to be more variable, at least on a relative basis.
15
16 E.2.6 Human Exposure to PM
17 In community epidemiology studies of PM and other air pollutants, ambient concentrations
18 are normally used as surrogates for personal exposure to pollutants of ambient origin. Since
19 people spend most of their time indoors, and the indoor environment is protective for most
20 ambient pollutants, it is important to understand the relationship between concentrations of
21 ambient pollutants measured at community monitoring sites and the contributions of those
22 concentrations to personal exposure. This is best done by considering separately (a) the
23 relationship between concentrations at a community air-monitoring sites(s) and immediately
24 outside an indoor environment, (b) the relationship between concentrations outside the indoor
25 environment and the contributions of the outdoor pollutant to the indoor environment, and (c) the
26 effect of activity patterns, i.e., time spent outdoors and in various indoor environments.
27
28 • Analyses of recent data from the PM25 monitoring network show reasonable site-to-site
29 correlation in most cities studied over distances of 20 to 50 km. This indicates that in such
30 cities the concentration at a community air-monitoring site (or the average of several such sites)
31 will provide an adequate representation of the concentration outside a home. Less information
April 2002 E-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 is available regarding site-to-site correlations for PM10_2 5, chemical components of PM (other
2 than sulfate which has high site-to-site correlations), and contributions from specific source
3 categories such as vehicular traffic-related PM. Even though site-to-site correlations may be
4 high, annual or seasonal averages may show sizable differences.
5
6 • The relationship between outdoor air pollutant concentrations and indoor concentrations due to
7 the concentration outdoors depends on the penetration factor (the fraction of the outdoor
8 concentration which reaches the indoor environment), how rapidly the indoor air is diluted by
9 outdoor air (measured as the air exchange rate), and the rate at which the ambient pollutant is
10 deposited or removed in the indoor environment. The deposition rate for PM is highly
11 dependent on particle size, being high for coarse and ultrafme particles but low for particles in
12 the accumulation-mode size range (0.1 to 1.0 //m diameter). Thus, the infiltration factor (the
13 ratio of the indoor concentration to the outdoor concentration) will be high for accumulation-
14 mode particles and for PM25 since most of the PM25 mass will be in the accumulation mode.
15
16 • Exposure also depends on the amount of time people spend outdoors. The attenuation factor,
17 in the case of PM, is defined as the ratio of the ambient PM exposure to the ambient PM
18 concentration, and accounts for the difference in the time spent indoors and outdoors as well as
19 the difference in exposure between indoors and outdoors.
20
21 • People are also exposed to particles and other pollutants generated indoors. It is not possible to
22 measure ambient PM exposure directly; only the combination of ambient and nonambient PM
23 exposure (total personal exposure to PM) can be measured. Ambient PM exposure must be
24 inferred or estimated from measurements of ambient concentration and total personal exposure.
25
26 • Major indoor sources are smoking, other indoor combustion, cooking, cleaning, and general
27 movement of people. Indoor particles are generated primarily in the ultrafme or coarse modes
28 and therefore have shorter indoor lifetimes than ambient-infiltrated particles (particles that have
29 penetrated indoors and remained suspended). The concentration of PM from indoor sources
30 appears to be independent of ambient concentrations, since personal activities generally do not
31 depend on ambient concentrations; however, this may change as more people are alerted to
April 2002 E-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 high pollution days and stay indoors. If nonambient PM exposure is independent (not
2 correlated with) ambient concentrations, a regression of measured personal exposures against
3 ambient concentrations will provide the average attenuation coefficient (slope of the regression
4 line) and the average nonambient concentration (the intercept). Such average values have been
5 obtained in several studies.
6
7 • It is more difficult to estimate individual daily values of ambient PM exposures. This could be
8 done for the PTEAM study because, in addition to ambient concentrations and personal
9 exposures, data were available on air exchange rates and time outdoors and the penetration
10 factor and deposition rate were estimated statistically. Ambient PM exposures can also be
11 estimated by using the personal sulfate/ambient sulfate ratio as an estimate of the attenuation
12 factor for PM2 5. This technique assumes that there are both minimal indoor sources of sulfate
13 and that the PM2 5 and sulfate have similar particle size distributions.
14
15 • Most exposure studies measure one (or a few) subjects on one day and a different one (or a
16 few) subjects on a different day. The highly variable nonambient exposure for different people
17 results in a low correlation between ambient concentration and total personal exposure for this
18 "pooled" data set. If a set of individuals each have their total personal exposure measured for
19 enough days to provide a meaningful relationship, it is observed that some of them will have
20 high correlations between ambient concentration and total personal exposure. The median
21 correlations from such studies ("longitudinal") are higher than that for the "pooled" data set.
22 If enough people are measured each day so that a meaningful daily average can be obtained, the
23 correlation between ambient concentration and the daily average community PM exposure is
24 high. Also, the correlation between ambient concentration and ambient PM exposure is high.
25 Therefore, ambient PM concentration appears to provide an adequate indicator of ambient PM
26 exposure for use in PM epidemiology studies, but such studies do not provide information on
27 the health effects of nonambient pollution (i.e., indoor-generated pollution).
28
29 • As long as the nonambient PM exposure is not correlated with the ambient PM exposure, it
30 will not bias the estimated health effect of PM. However, the effect per //g/ambient PM
31 concentration will be biased low compared to the health effect per //g/ambient PM exposure by
April 2002 E-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 the attenuation factor. This effect probably explains some of the heterogeneity in PM10 effects
2 observed in multicity epidemiology studies, as indicated by a correlation of PM effects in
3 different cities with air conditioning use in those cities (i.e., the higher the air conditioning use,
4 the lower the health effect estimate per //g/m3 of ambient PM).
5
6 • Exposure relationships also provide some insight into the issue of confounding. While the data
7 base is small, concentrations of gaseous co-pollutants, NO2, O3, and SO2 (and probably CO) are
8 likely poorly correlated, and sometimes not significantly correlated, with personal exposure to
9 the respective co-pollutant. However, they are frequently significantly correlated with both the
10 ambient PM concentration and the ambient PM exposure. Thus, in a regression, where
11 associations are found between gaseous co-pollutants and a health effect, it may be because
12 they are a surrogate for PM rather than a confounder. That is, the health effect due to PM is
13 transferred to the gaseous pollutant because of the positive correlation between the ambient
14 concentration of the gas and the ambient PM exposure.
15
16
17 E.3 DOSIMETRY
18 Knowledge of the dose of particles delivered to a target site or sites in the respiratory tract
19 is important for understanding possible health effects associated with human exposure to ambient
20 PM and for extrapolating and interpreting toxicologic data obtained from studies of laboratory
21 animals. Particles of different sizes are subject to large differences in regional respiratory tract
22 deposition, translocation, clearance mechanisms and pathways, and consequent retention times.
23 Key findings derived from the assessment of dosimetry information include:
24
25 • Respiratory tract deposition patterns are dependent on particle size, as indicated by the
26 aerodynamic or thermodynamic diameter of the particles within the inspired air. Biologic
27 effects may be a function not only of particle mass deposition but also of particle number; the
28 total surface area of the particles; or the acidity, surface chemistry, or charge of the particles.
29
30 • Particles may be deposited in the extrathoracic (ET) region (i.e., mouth, nose, pharynx, and
31 larynx); the conducting airways of the tracheobronchial (TB) region; and the alveolar (A)
April 2002 E-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 region, where gas exchange occurs. There are differences in deposition mechanisms and dose
2 distribution in each of these regions that are dependent on the physical characteristics of
3 particles and on airway geometry.
4
5 • Particles deposit in the respiratory tract mainly by five mechanisms: (1) inertial impaction,
6 (2) sedimentation, (3) diffusion, (4) electrostatic precipitation, and (5) interception. Impaction
7 is an important deposition mechanism for particles >l//m in large extra- and intrathoracic
8 airways at higher flows; sedimentation and diffusion are more important for particles >0.5 //m
9 and <0.3 //m, respectively, at low flow rates in smaller airways. Particles between 0.3 and
10 0.5 //m in size are small enough to be little influenced by impaction or sedimentation and large
11 enough to be minimally influenced by diffusion; and, so, they undergo the least respiratory tract
12 deposition. Electrostatic precipitation is deposition related to particle charge; effects of charge
13 on deposition are inversely proportional to particle size and airflow rate. The interception
14 potential of any particle depends on its physical size rather than its aerodynamic size.
15
16 • Hygroscopicity, the propensity of a material for taking up and retaining moisture, is a property
17 of some ambient particle species and affects respiratory tract deposition. Hygroscopicity
18 generally increases deposition in the TB and A regions for particles with initial sizes larger than
19 -0.5 //m or smaller than «0.01, but decreases deposition for intermediate sizes.
20
21 • The ET region acts as an efficient filter that reduces penetration of inhaled particles to the TB
22 and A regions of the lower respiratory tract. Total respiratory tract deposition increases with
23 particle size for particles >1.0 //m, is at a minimum for particles 0.3 to 0.5 //m, and increases as
24 particle size decreases below that range.
25
26 • Enhanced particle retention occurs on carinal ridges in the trachea and through segmental
27 bronchi; and deposition "hot spots" occur at airway bifurcations or branching points. Peak
28 deposition sites shift from distal to proximal sites as a function of particle size, with greater
29 surface dose in conducting airways than in the A region for all particle sizes. However, surface
30 number dose (particles/cm2/day) is much higher for fine particles than for coarse for typical
31 bi-modal ambient aerosols.
April 2002 E-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Extrathoracic deposition of ultrafine particles (<0.1 //m) is very high; as particle size decreases
2 below 0.1 (j,m, particles tend to behave more like gases. Estimates of extrathoracic deposition
3 range from 50% for oral breathing to >90% for nasal breathing. Within the thoracic region, the
4 deposition distribution of ultrafine particles is highly skewed towards the proximal airway
5 regions and resembles the deposition of coarse particles.
6
7 • Various host factors have been shown to influence particle deposition patterns, including
8 airway dimensions (size and shape), breathing pattern (flow and volume), and the presence of
9 obstructive or inflammatory airway disease. The ET deposition is higher with nose breathing
10 than for mouth breathing, with increased ventilation rates associated with increasing levels of
11 physical activity or exercise leading to more oronasal breathing and increased delivery of
12 inhaled particles to TB and A regions in the lung. Gender and age differences in the
13 homogeneity of deposition, as well as deposition rate, could affect susceptibility. Children, for
14 example, would receive greater doses of particles per lung surface area than would adults.
15 Also, obstructive airway diseases (such as asthma and chronic bronchitis) result in increased
16 deposition of particles in the central airway region and distal lung regions receiving greater
17 ventilation.
18
19 • Particles depositing on airway surfaces may be cleared from the respiratory tract completely or
20 translocated to other sites within this system by regionally specific clearance mechanisms.
21 Clearance is either absorptive (dissolution) or nonabsorptive (transport of intact particles).
22 Deposited particles may be dissolved in body fluids, taken up by phagocytic cells, or
23 transported by the mucociliary system. Retained particles tend to be small (<2.5 //m) and
24 poorly soluble (e.g., silica, metals).
25
26 • Tracheobronchial clearance has both a fast and a slow component. In the fast phase particles
27 deposited in the TB region clear out rapidly during the first several hours and continue to clear
28 out for 24 hours. A small remaining portion may clear out over several days (slow phase).
29 Translocation of poorly soluble PM to the lymph nodes takes a few days and is more rapid for
30 smaller (< 2 //m) particles; elimination rates of these retained particles are on the order of
31 years. People with COPD have increased particle retention partly because of increased initial
April 2002 E-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 deposition and impaired mucociliary clearance and use cough to augment mucociliary
2 clearance.
3
4 • Alveolar clearance takes months and years. Particles may be taken up by alveolar macrophages
5 within 24 hours, but some phagocytosed macrophages translocate into the interstitium or
6 lymphatics whereas some remain on the alveolar surface. Penetration of uningested particles
7 into the interstitium increases with increasing particle load and results in increased
8 translocation to lymph nodes.
9
10 • Acute effects of PM are probably best related to deposited dose, whereas chronic effects may
11 be related to cumulative or retained dose. Retention of particles is a function of deposition site,
12 clearance of particles by macrophages or the mucociliary system, and particle characteristics,
13 especially solubility. Chronic effects may also arise from recurring cycles of pulmonary injury
14 and repair.
15
16 • Mathematical and computational fluid dynamic models are available to predict deposition,
17 clearance, and retention of particles in the respiratory tract. Although these models have
18 become more sophisticated and versatile, validation of the models is still needed.
19
20 • A better understanding of species differences in deposition, translocation, and clearance of
21 particles, especially ultrafme particles, is still needed. So are better models of extrapolation
22 between animals used in inhalation studies and humans.
23
24
25 E.4 PARTICULATE MATTER HEALTH EFFECTS
26 E.4.1 Toxicology of Particulate Matter in Humans and Laboratory Animals
27 Toxicological research on ambient PM or combustion-related particles is used to address
28 several related questions that are important toward an understanding of the cardiopulmonary
29 effects that have been reported in PM-exposed human populations.
30
April 2002 E-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Does exposure to PM at relevant ambient concentrations cause toxicological effects?
2 • What characteristics of PM contribute to the observed toxicity?
3 • What factors affect individual or subpopulation susceptibility to the effects of PM?
4 • What are the combined effects of PM and other pollutants in the ambient air?
5 • What mechanisms may be involved in the toxicological response to PM exposure?
6
7 Data on the toxicology of PM are derived from controlled inhalation exposure studies of
8 humans and laboratory animals, intratracheal instillation studies in humans and animals, and
9 ex vivo studies of human and animal cells grown in culture. The human or animal populations
10 (cells) studied vary by age, health status, or other host factors. As seen in the previous section,
11 deposition of PM in the respiratory tract depends on particle size and regional distribution.
12 Potential biologic effects may be a function not only of particle mass deposition but also of
13 particle number, the total surface area of the particles, particle acidity, and the surface chemistry,
14 charge, and composition of the particle in addition to other exposure variables (e.g., duration,
15 temperature, humidity, activity levels). Responses to PM in the respiratory tract also are
16 dependent on the physiological status of the host, as well as the translocation of PM or PM
17 constituents to other sites. Ex vivo studies provide important additional information regarding
18 mechanisms of action of PM or PM constituents on cells or cellular components.
19 The data available in the previous 1996 PM AQCD and in other published documents were
20 mainly from studies that investigated the respiratory effects of specific components of
21 combustion-related particles from mobile or stationary sources (e.g., diesel particles, fly ash),
22 ambient particles, or laboratory-derived surrogate particles (e.g., sulfuric acid droplets). In this
23 document, more emphasis is placed on assessment of new data obtained from controlled studies
24 of particles collected from emission sources or ambient samplers (e.g., impactors, diffusion
25 denuders) and by the use of aerosol concentrators that provide a technique for exposing humans
26 or laboratory animals by inhalation to concentrated ambient particles (CAPs). Key findings
27 derived from the assessment of these effects include:
28
29 • Combustion-related particles (fly ash and urban air particles) from a large number of emission
30 sources and ambient airsheds cause a spectrum of responses in the airways of laboratory
31 animals and humans. These include inflammation, cellular injury, and increased permeability.
April 2002 E-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 Soluble metal components (e.g., Cu, Fe, Ni, V, Zn) of combustion particles have been
2 implicated in the responses, possibly related to oxidant production and release of intercellular
3 signaling molecules (cytokines).
4
5 • Toxicological studies of aqueous extracts from ambient PM collected on filters in the Utah
6 Valley around Provo, UT demonstrated increased pulmonary inflammatory effects after airway
7 instillation exposure of humans and laboratory animals, and after direct exposure to cells in
8 culture. Extract analysis of particle components acquired during operation of an open-hearth
9 steel mill identified more sulfate, cationic salts (Ca, K, Mg), and metals (As, Cu, Fe, Mn, Ni,
10 Pb, Sr, Zn). The inflammatory response was significantly reduced when the steel mill was
11 closed, thus tending to corroborate epidemiology findings for the same time period, indicating
12 that Utah Valley residents reported decreased hospital admissions for respiratory diseases.
13
14 • Cells primed by inflammatory mediators show increased cytokine responses to PM.
15 Combustion-related particles may cause increased oxidant production, presumably related to
16 metal components of particles, and damage to cells in vitro. Responses include impaired
17 macrophage phagocytosis and altered permeability.
18
19 • Acute exposures to soluble transition metals can cause inflammatory responses in the
20 respiratory tract of humans and laboratory animals. The effective exposure levels (mg/m3) are
21 typically much higher than typical ambient air metal concentrations (<15 |ig/m3) in the U.S.
22 atmosphere.
23
24 • Endotoxin, a lipopolysacharide associated with bacteria, and a common contaminant of
25 ambient PM, also causes inflammation in humans and laboratory animals at concentrations
26 (>0.5 jig) that are much higher than typically found in the ambient air (<0.5 ng/m3).
27
28 • Human inhalation exposure to diesel exhaust particles causes increased acute sensory and
29 respiratory symptoms, lung inflammation, and impairment of alveolar macrophage function.
30 Effects in laboratory animals include pulmonary histopathology and chronic inflammation.
31 These noncancer effects are thought to be due to the organic carbon constituents or to
April 2002 E-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 metabolites of metal components of the particle. Except for diesel exhaust particles, no other
2 research has been published on acute effects of organic carbon constituents which often
3 comprise a substantial portion of ambient PM (10 to 60% of the total dry mass).
4
5 • Recent studies report systemic changes in rodents and dogs exposed to high concentrations of
6 instilled and inhaled ambient PM and combustion-related particles, including alteration of heart
7 rate (e.g., bradycardia, arrhythmia) and blood pressure, hypothermia, alterations in blood cells,
8 and increased blood levels of endothelins (vasoactive substances) and fibrinogen (coagulation
9 factors).
10
11 Mixtures
12 • Mixtures of ozone and PM (e.g., urban PM, diesel PM, sulfate aerosols, ultrafine carbon) may
13 cause enhanced effects on lung cells, increased inflammation, and decrements in human lung
14 function. In one controlled human study, a mixture of ozone and CAPs produced peripheral
15 vasoconstriction, possibly caused by autonomic reflexes or as a result of increased circulating
16 endothelins.
17
18 Mechanisms
19 • A number of studies indicate that increased production of inflammatory cytokines and reactive
20 oxidant species (ROS) may play a role in PM-induced responses. The cytokine responses
21 correlate with endotoxin, which is an important component of ambient coarse- and fine-mode
22 PM. Catalysis of ROS is likely related to soluble metals in ambient PM and combustion-
23 related particles.
24
25 • Somatosensory neurons of the autonomic nervous system (ANS) may also be affected by the
26 inflammatory response to ambient PM, especially when there is epithelial airway damage.
27 Pulmonary reflex responses through the ANS can have direct effects on the heart and may
28 cause other systemic effects.
29
30 • Studies on ultrafine compared to fine-mode particles indicate a greater response to ultrafine
31 particles in regards to airway inflammation, an effect that appears to be related to their greater
April 2002 E-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 surface area. There also is some evidence to suggest that ultrafine PM may exit the lung and
2 deposit in other organs, including the heart and liver.
3
4 • Combustion-related particles (e.g., oil fly ash) and urban PM can induce apoptosis
5 (programmed cell death) of human alveolar macrophages.
6
7 • Other potential cellular and molecular mechanisms include changes in the expression of
8 specific growth factors, adhesion molecules, stress proteins, matrix proteinases, transcription
9 factors, and antioxidant enzymes.
10
11 Susceptibility
12 • Chemically or pharmacologically treated rat "models" of cardiopulmonary disease (e.g.,
13 hypertension, chronic bronchitis), as well as older rats, manifest increased cardiopulmonary
14 responses, lung damage, and even death after exposure to ambient PM and combustion-related
15 particles by intratracheal instillation or inhalation.
16
17 • Inhalation or instillation of combustion-related particles (e.g., oil fly ash) and diesel exhaust
18 PM may augment the immune response to antigens in allergic animals or humans. These
19 studies provide a plausible mechanism for an association between combustion-source PM
20 exposure and exacerbation of asthma.
21
22 • Acid aerosols cause little or no changes in pulmonary function in healthy subjects; however,
23 asthmatics may develop small, but potentially relevant increased airway responsiveness. New
24 information relating acid aerosol exposure to cardiovascular effects in laboratory animals is
25 interesting, but needs further investigation.
26
27 • Genetic susceptibility can plan a role in the response to inhaled or instilled particles.
28
29 E.4.2 Population Groups at Risk
30 Susceptibility can be affected by factors which influence dosimetry or the response of
31 tissues to particle burdens. Host factors that may increase the susceptibility to PM include both
April 2002 E-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 changes in physiologic factors affecting respiratory tract deposition and pathophysiologic factors
2 affecting response.
3
4 • Susceptible groups most clearly at special risk for PM effects include the elderly and those with
5 cardiopulmonary disease, based on available epidemiology findings.
6
7 • Epidemiology findings indicate that risk of mortality and morbidity due to lower respiratory
8 disease (e.g. pneumonia) is increased by ambient PM exposure. This may be due to
9 exacerbation by PM of already existing respiratory disease. PM may also increase
10 susceptibility to infectious disease by decreasing clearance, impairing macrophage function, or
11 through other specific and nonspecific effects on the immune system. The epidemiologic
12 findings also indicate that individuals with preexisting infectious respiratory disease (e.g.,
13 pneumonia) are at increased risk for PM effects.
14
15 • Epidemiologic findings indicate that ambient PM exposures are also associated with increased
16 risk for mortality and hospitalization due to cardiovascular causes. Cardiac arrhythmia has
17 been hypothesized as being involved in mortality due to acute PM exposure. Thus, individuals
18 with pre-existing cardiovascular disease(s) are likely a susceptible group at increased risk for
19 ambient PM effects
20
21 • Studies of infants and children indicate that they are a potentially susceptible population.
22 Panel studies on asthma and other respiratory conditions show exacerbation by PM exposure.
23 Children are susceptible to respiratory effects associated with PM exposure from pre-natal and
24 post-natal effects through exacerbation of asthma and respiratory symptoms in school age
25 children.
26
27 E.4.3 Epidemiology Findings
28 Epidemiologic evidence concerning the mortality and morbidity effects of ambient PM has
29 expanded greatly since the 1996 PM Air Quality Criteria Document (PM AQCD). The most
30 important enhancements in information include:
31
April 2002 E-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 • New multi-city studies of health endpoints using ambient PM10 mass concentrations and/or
2 other ambient PM indicators (e.g. Black Smoke, Coefficient of Haze, etc).
3
4 • New studies on a variety of endpoints using ambient fine particle fraction (PM2 5), a limited
5 number using ambient coarse thoracic PM fraction (PM10_2 5), and a few using ambient ultrafine
6 particle (PMai) mass concentrations;
7
8 • New studies in which the relationship of some health endpoints to ambient particle number
9 concentrations were evaluated.
10
11 • Additional studies which evaluated the sensitivity of estimated PM effects to the inclusion of
12 gaseous co-pollutants in the model.
13
14 • New studies evaluating the effects of specific source categories of air pollution based on
15 empirical combinations (factor analysis).
16
17 • Further studies of cardiopulmonary endpoints associated with PM exposures. Cardiovascular,
18 as well as respiratory, causes of death and hospitalization in older adults may be a significant
19 component of PM-attributable mortality.
20
21 • New studies suggest that infants and children may represent an additional subgroup at special
22 risk for ambient PM exposure effects. The new results most clearly indicate that children
23 appear to be susceptible to respiratory effects associated with ambient PM exposures, including
24 exacerbation of asthma and respiratory symptoms in school-age children.
25
26 • A few studies also report ambient PM to be associated with intrauterine growth reduction and
27 low birth weight (known infant health risk factors) and excess infant mortality. However, no
28 toxicologic evidence has yet been advanced to support biological plausibility of such effects
29 due to ambient PM or to identify the pathophysiologic mechanisms involved.
30
April 2002 E-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 • New analyses of American Cancer Society (ACS) data extended over more years and
2 addressing criticisms of prior ACS analyses not only substantiate previous ACS study findings
3 of PM associations with increased risk of cardiopulmonary-related mortality/morbidity but also
4 find PM associations with increased risk of lung cancer.
5
6 • PM health effects have been reported to be associated with several different ambient PM size
7 fractions (ultrafine, fine, coarse); individual chemical components (sulfate, nitrate, elemental
8 carbon, organic carbon) and specific source categories (vehicular traffic, regional sulfate,
9 vegetative burning, and fuel oil combustion).
10
11 • Various health effects may occur at different time scales for exposure to PM, from
12 (a) short -term responses to hourly to daily exposures through (b) larger excess mortality
13 associated with medium-term exposures (15 to 120 day averages) to (c) excess morbidity or
14 mortality associated with long-term (multi-year) exposures.
15
16 • Because PM indicators and the gaseous co-pollutants, CO, NO2, SO2, and O3, are frequently
17 significantly correlated, the potential exists for the confounding of the adverse health effects
18 attributable to PM25 in short-term exposure studies by exposure to gaseous co-pollutants. This
19 makes it difficult to apportion the risk among PM acting alone, PM acting in combination with
20 gaseous co-pollutants, the gaseous co-pollutants per se, a specific source category, or the
21 overall ambient pollutant mix. However, recent exposure studies suggest that the ambient
22 concentrations of the gaseous co-pollutants, although frequently correlated with ambient PM
23 concentrations, are not well correlated with the personal exposure to the respective gaseous
24 co-pollutants. Therefore, the gaseous co-pollutants are not likely to be confounders, rather they
25 are likely surrogates for PM (or specific source categories such as vehicular-traffic-related
26 particles or regional sulfate). The low exposures to the reactive gaseous co-pollutants (NO2,
27 SO2, O3) as well as to CO, of people who spend most of their time indoors, relative to known
28 toxic levels, also suggests that these gaseous pollutants are unlikely to be responsible for the
29 health effects found to be associated with PM (although the gaseous pollutants may also
30 independently exert effects on health, as well).
31
April 2002 E-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 E.4.3.1 Ambient PM Mortality Effects
2 • New multi-city studies convincingly demonstrate the association of PM10 with health effects.
3 The use of the same statistical model for many cities with different PM concentrations and
4 compositions, and different correlations of PM and gaseous co-pollutants, strengthens the
5 reliability of the associations and also demonstrates that the effects of PM are not confounded
6 by the gaseous co-pollutants.
7
8 • Several studies have considered PM2 5 and PM10_2 5 as well as PM10. In same cases PM2 5, and in
9 some cases PM10_2 5, was more significant than PM10 and had a higher excess risk per //g/m3
10 increase in PM. One study found both PM2 5 and PM10_25 to be statistically significant and to
11 have excess risks that were higher than PM10 and that peaked on different lag days. Thus,
12 epidemiologic studies also support the separation of PM10 into fine and thoracic coarse fraction
13 components. However, there are some suggestions that the toxicity of PM10_2 5 may not be due
14 to the crustal or soil components per se, but rather more likely to toxic trace metals or organic
15 compounds carried into the lungs on the coarse particles.
16
17 • Source apportionment techniques have been used to obtain groupings of PM components and
18 gaseous co-pollutants (factors) with minimal correlation among factors. Use of these factors
19 permits determination of associations with health effects with little potential for confounding
20 among the factors. These factors can frequently be associated with specific source categories.
21 However, the association of factors with source categories may be subjective and a factor may
22 contain contributions from more than one source category.
23
24 • A vehicular traffic related factor has been identified in all four studies that examined that
25 factor. Although not all studies measured all species, this factor appears to contain PM2 5, CO,
26 NO2, EC, and OC, as well as specific elements Mn, Fe, Zn, and Pb, that might be emitted or
27 resuspended by traffic. Epidemiology alone cannot apportion the health effect among these
28 different components of the vehicular traffic related factor. However, the low potential for
29 confounding of PM10 by the CO and NO2 suggests that they may serve as surrogates for the PM
30 component of the vehicular traffic-related factor.
31
April 2002 E-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Independent associations, not confounded by gaseous co-pollutants have been found for nitrate
2 (by single and multiple regression with gaseous co-pollutants) and for regional sulfate and a
3 PM factor associated with vegetative burning (by use of source category factors). Thus, the
4 major components of PM25 by mass, sulfate, nitrate, and organic, have been found to have
5 significant associations with mortality.
6
7 • A major concern has been that the effects attributed to PM may really be due to one or more of
8 the gaseous co-pollutants (CO, NO2, SO2, O3). Epidemiology cannot demonstrate causality,
9 only association. However, multi-city studies in conjunction with exposure relationships have
10 been able to demonstrate that the gaseous co-pollutants are more likely to be surrogates than
11 confounders. This does not demonstrate that the gaseous co-pollutants do not have real effects,
12 only that they are significantly correlated with the PM indicator but not with the personal
13 exposure to the gaseous co-pollutants.
14
15 • The results of long-term exposure studies, indicative of increased mortality and/or morbidity
16 risks being associated with exposures to ambient PM over many years, have been substantiated
17 both by independent reanalyses and updated, extended analyses of more years of data and cases
18 of mortality or morbidity. This includes much stronger evidence for ambient PM effects on
19 cardiovascular and respiratory endpoints, as well as strong evidence for PM-related increases in
20 lung cancer risks.
21
22 E.4.3.2 Ambient PM Morbidity Effects
23 Numerous epidemiologic studies in the United States and elsewhere have also
24 demonstrated significant associations between ambient PM exposures indexed by a variety of
25 indicators (PM10, PM25, PM10.25, SOJ, H+, BS, COH, TSP) and various acute and chronic
26 morbidity outcomes. Such outcomes include, for example, hospital admissions, medical visits,
27 increased respiratory symptoms, and decreased lung function.
28
29 • The ecologic time series studies add substantially to the body of available literature for effects
30 of PM10 on acute CVD hospital admissions. Results for adult cardiovascular mortality are
31 qualitatively consistent with those for hospital admissions. However, uncertainties regarding
April 2002 E-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 the possible role of co-pollutants complicates interpretation with respect to independent PM10
2 effects. New studies evaluating longitudinal association between ambient PM and
3 physiological measures of cardiovascular function or biochemical changes in the blood that
4 may be associated with cardiac risk present a range of findings suggesting possible adverse
5 effects of PM on cardiac rhythm and other biological functions.
6
7 • The results of new studies are generally consistent with regard to ambient PM associations of
8 short-term exposures with respiratory-related hospital admissions/medical visits. The excess
9 risk estimates fall most consistently in the range of 5 to 25% per 50 //g/m3 PM10 increments,
10 with those for asthma visits and hospital admissions tending to be somewhat higher than for
11 COPD and pneumonia hospital admissions.
12
13 • The peak flow analyses results for asthmatics tend to show small decrements for both PM10
14 and PM2 5. The effects of PM on respiratory symptoms in asthmatics tended to be positive,
15 although they were much less consistent than the effects on lung function. Most PM10 studies
16 showed increases in cough, phlegm, difficulty breathing, and bronchodilator use, although
17 these increases were generally not statistically significant.
18
19 • Results of the PM10 peak flow analyses in non-asthmatic studies were inconsistent, with fewer
20 studies reporting results in the same manner as for the asthmatic studies. The effects on
21 respiratory symptoms in non-asthmatics were similar to those in asthmatics. Most studies
22 showed that PM10 increases cough, phlegm, difficulty breathing, and bronchodilator use,
23 although these increases were generally not statistically significant.
24
25 • Differences in peak flow and bronchitis symptoms and prevalence rates in children were found
26 to be somewhat more closely associated with annual average H+ concentrations than with other
27 PM indicators. However, in studies demonstrating these effects, the acid levels were highly
28 correlated with other fine-particle indicators.
29
30 • While numerous studies of PM related respiratory morbidity have been conducted using PM10
31 as an indicator, only a few studies have examined the effects of fine and coarse fraction particle
April 2002 E-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 indicators separately. The PM2 5 studies do show effects related to exposure to the fine fraction,
2 but high correlations among PM2 5, PM10, and acid aerosols make it very difficult to attribute
3 the effects to a single specific exposure indicator. A few studies also indicate that respiratory
4 morbidity effects (increased respiratory symptoms) are associated with ambient coarse fraction
5 (PM10_2 5) concentrations.
6
7
8 E.5 ENVIRONMENTAL EFFECTS OF PM
9 E.5.1 Vegetation and Ecosystem Effects
10 • Deposition of particulate matter from the atmosphere has the potential to alter ecosystem
11 structure and function. Human existence on this planet depends on the essential life-support
12 services that ecosystem structure and functions provide. Concern has risen in recent years
13 regarding the consequences of changing the biological diversity of ecosystems because human
14 activities are creating disturbances that are altering the structure (complexity and stability) and
15 functioning (producing changes in energy and water flow and nutrient cycling) of ecosystems.
16
17 • Human-induced changes in biotic diversity and alterations in the structure and functioning of
18 ecosystems are the two most dramatic ecological trends in the past century. Biodiversity
19 encompasses all levels of biological organization, including individuals, populations, species,
20 and ecosystems. For this reason, there is a need to understand the effects of PM deposition on
21 vegetation and ecosystems and biodiversity.
22
23 • Ecosystem functions maintain clean water, pure air, a green earth (biodiversity) and impart the
24 following benefits: fixation of solar energy, absorption and breakdown of pollutants, cycling of
25 nutrients, binding of soil, degradation of organic wastes, maintenance of a balance of
26 atmospheric gases, regulation of radiation balance, and climate.
27
28 • The relationship between PM and effects on vegetation and ecosystems is dependent on the
29 size, origin, and chemical constituents of the particles. Exposure to a given mass concentration
April 2002 E-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 of PM may, depending on the particular mix of deposited particles, lead to widely differing
2 phytotoxic responses.
3
4 • Atmospheric deposition of particles to ecosystems takes place via both wet and dry processes
5 through the three major routes indicated below:
6 (1) Precipitation scavenging in which particles are deposited in rain and snow
7 (2) Fog, cloud-water, and mist interception
8 (3) Dry deposition, a much slower, yet more continuous removal to surfaces.
9
10 • Deposition of heavy metal particles to ecosystems occurs by wet and dry processes. Dry
11 deposition is considered more effective for coarse particles of natural origin and elements such
12 as iron and manganese, whereas wet deposition generally is more effective for fine particles of
13 atmospheric origin and elements such as cadmium, chromium, lead, nickel, and vanadium.
14
15 • The actual importance of wet versus dry deposition, however, is highly variable, depending on
16 the type of ecosystem, location and elevation. The range of particle sizes, the variety of
17 chemical constituents in airborne PM, and the diversity of canopy surfaces, have slowed
18 progress in both prediction and measurement of dry particulate deposition. Wet deposition
19 generally is confounded by fewer factors and has been easier to quantify.
20
21 • Parti culate matter, when transferred from the atmosphere to plant surfaces, may cause direct
22 effects when it (1) resides on the leaf, twig or bark surface for an extended period; (2) is taken
23 up through the leaf surface; or produce indirect effects when (3) removed from the plant via
24 suspension to the atmosphere, washing by rainfall, or by litter-fall with subsequent transfer to
25 the soil.
26
27 • Deposition of PM on above-ground plant parts can have either a physical and or chemical
28 impact, or both. The effects of "inert" PM are mainly physical, while the effects of toxic
29 particles are both chemical and physical. The majority of the easily identified direct and
30 indirect effects occur in severely polluted areas around heavily industrialized point sources
31 (such as limestone quarries, cement kilns, iron, lead, and various smelting factories).
April 2002 E-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Chemical effects include excessive alkalinity or acidity. Studies of the chemical additions of
2 PM to foliage have found little or no effects on foliar processes unless exposure levels were
3 significantly greater than would typically be expected in the ambient environment. The effects
4 of dust deposited on plant surfaces or on soil are more likely to be associated with their
5 chemistry than with the mass of deposited particles and are usually of more importance than
6 any physical effects. The effects of limestone dust on plants can cause long-term changes in the
7 structure, composition and function of the seedling-shrub and sapling strata of ecosystems in
8 the vicinity of limestone quarries and processing plants.
9
10 • Secondary organics formed in the atmosphere and referred to as toxic substances, persistent
11 organic pollutants (POPS), pesticides, hazardous air pollutants (HAPS), air toxics, and/or
12 semivolatile organic compounds (SOCS), are chemical substances not controlled by National
13 Ambient Air Quality Standards under Sect. 109 of the Clean Air Act (U.S. Code, 1994), but
14 rather are controlled under Sect. 112, Hazardous Air Pollutants. Mention of them is made in
15 this document because many form or attach to particles in the atmosphere. As particles they
16 become airborne and can be distributed over a wide area and impact remote ecosystems. Some
17 are of concern to humans because they may reach toxic levels in food chains of animals as well
18 as humans; others tend to decrease or maintain the same toxicity as they move through the food
19 chain.
20
21 • The depletion of stratospheric ozone caused by the release of chloroflurocarbons (CFC's) and
22 substances such as halides has resulted in heightened concern about potentially serious
23 increases of solar UV-B (SUVB) reaching the earth's surface. Terrestrial vegetation is
24 vulnerable to UV-B because of the need for sunlight during photosynthesis. Effects of UV-B
25 on plant growth are likely to be incremental. However, plants grown in full sunlight (because
26 they evolved under ambient UV-B radiation and have developed adaptive mechanisms) are not
27 as sensitive as plants gown under weak visible light. Therefore, plant species vary enormously
28 in their responses to UV-B exposures. In addition, large differences in response occur among
29 genotypes within a species and dicotyledons are more sensitive than monocotyledons.
30
April 2002 E-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Photosynthetic processes underlie the contributions of vegetation to nutrient cycling and energy
2 flow. Light penetration into plant canopies limits photosynthetic productivity. Regional haze
3 has been estimated to diminish surface visible solar radiation by approximately 8%.
4 An assessment of the direct effects on crop production suggests that optimal crop yields are
5 significantly affected by regional haze.
6
7 • Most PM deposited on vegetation eventually enters the soil environment, which is one of the
8 most dynamic sites of biological interaction. The major impacts on vegetation and ecosystems
9 are the indirect effects that occur through the soil and affect plant growth, vigor and
10 reproduction. Changes in nutrient cycling and plant nutrient uptake determine plant and
11 ecosystem responses.
12
13 • Bacteria and fungi in the soil have an important role in plant nutrition. Bacteria are essential
14 components of the nitrogen and sulfur cycles that make these elements available for plant
15 uptake. Fungi form mycorrhizae, a mutualistic symbiotic relationship with plant roots that is
16 integral to the uptake of mineral nutrients. The impact of nitrates, sulfates and metals in PM is
17 determined by their affect on the growth and functions of the bacteria and fungi involved in
18 making nutrients available for plant uptake.
19
20 • Extensive evidence indicates that heavy metals deposited from the atmosphere to forests
21 accumulate either in the richly organic forest floor, where the biological activity is the greatest,
22 or in the soil layers immediately below. Accumulation of heavy metals in litter presents the
23 greatest potential for altering nutrient cycling. Increased amounts of litter in metal-
24 contaminated areas appear to result from reduced activity of microorganismal populations.
25
26 • Phytochelatins are intracellular metal-binding peptides that act as indicators of metal stress.
27 Because they are produced by plants as a response to sublethal concentrations of heavy metals,
28 they are indicators that heavy metals are involved in forest decline. Concentrations of
29 phytochelatins were observed to increased with altitude as did forest decline and they also
30 increased across the regions that showed increased levels of forest injury.
31
April 2002 E-29 DRAFT-DO NOT QUOTE OR CITE
-------
1 • The major impact of atmospheric PM on ecosystems is indirect and occurs in the soil
2 environment through the deposition of nitrates and sulfates and the acidifying effects of the
3 H+ ion associated with these compounds in wet and dry deposition. Most nitrate is not
4 deposited or measured as PM but is a combination of wet and dry deposition.
5
6 • The deposition of nitrates, and the acidifying effects of the H+ ion in association with nitrates
7 and sulfates in precipitation, causes soil acidification, changes the Ca/Al ratio, affects the
8 growth of soil bacteria and fungi, and alters biogeochemical cycling, all of which affects the
9 growth of natural vegetation and forest trees. Leaching of nitrates and other minerals through
10 runoff can also affect streams and coastal and aquatic wetlands and thus influence their ability
11 to produce the products and services necessary for human society.
12
13 • Intensive research over nearly a decade indicates that, although the soils of most North
14 American forests are nitrogen limited, severe symptoms of nitrogen saturation have been
15 observed in: high-elevation spruce-fir forests of the Appalachian Mountains, hardwood
16 watersheds near Parsons, WV; watersheds in the Los Angles Air Basin; high-elevation alpine
17 watersheds in the Colorado Front Range; and a deciduous forest in Ontario, Canada.
18
19 • Nitrogen saturation results in a progressive syndrome of concurrent responses to long-term,
20 chronic nitrogen deposition. As nitrogen reaches saturation in temperate-zone forests, there are
21 decreases in nitrogen mineralization and increases in trends of foliar Mg:N and Ca:Al ratios.
22 Preliminary evidence suggests some forests may decline in productivity and experience greater
23 mortality as a result of chronic nitrogen deposition.
24
25 • Increases in soil nitrogen play a selective role. Plant succession patterns and biodiversity in
26 some ecosystems are significantly affected by chronic nitrogen additions. Long-term nitrogen
27 fertilization studies in both New England and Europe suggest that forests receiving chronic
28 inputs of nitrogen may decline in productivity and experience greater mortality. Studies also
29 suggest that declining coniferous forest stands with slow nitrogen cycling may be replaced by
30 deciduous fast-growing forests which cycle nitrogen rapidly.
31
April 2002 E-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Excess nitrogen inputs to unmanaged heathlands in the Netherlands has resulted in nitrophilous
2 grass species replacing slower growing heath species. Over the past several decades the
3 composition of plants in the forest herb layers had been shifting toward species commonly
4 found in nitrogen-rich areas. It also was observed that the fruiting bodies of mycorrhizal fungi
5 had decreased in number, indicating that formation of mycorrhizae were affected.
6
7 • The effects of excessive deposition of nitrogen, particularly NH3 and NH4+, have led to changes
8 in Dutch heathlands via: (1) acidification of the soil and the loss of cations of K+, Ca2+ and
9 Mg2+; and (2) nitrogen enrichment, which results in increased plant growth rates and altered
10 competitive relationships. Alteration of any of a number of parameters (e.g., increased
11 nitrogen) can alter ecosystem structure and function.
12
13 • There is a major concern that soil acidification will result in nutrient deficiency. Growth of tree
14 species can be affected when high aluminum to nutrient ratios limit uptake of calcium and
15 magnesium. Calcium is essential in the formation of wood and the maintenance of cells, the
16 primary plant tissues necessary for tree growth. Calcium must be dissolved in soil water to be
17 taken up by plants. A major concern is that soil acidity will lead to nutrient deficiency.
18
19 • Acid deposition can increase the aluminum concentrations in soil water by lowering the pH in
20 aluminum-rich soils through dissolution and ion-exchange processes. Aluminum in soil can be
21 taken up by roots more readily than calcium because of its greater affinity for negatively
22 charged surfaces. Tree species can be adversely affected if high Ca/Al ratios impair Ca and
23 Mg uptake.
24
25 • Ecosystem processes and productivity, nitrogen mineralization rates, and nitrate leaching
26 respond directly to human modification of ecosystems and to changes in atmospheric
27 composition and climate.
28
29 E.5.2 Particulate Matter-Related Effects on Materials
30 Atmospheric PM and SO2 exert effects on materials that are related both to aesthetic appeal
31 and physical damage. Studies have demonstrated particles, primarily consisting of carbonaceous
April 2002 E-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 compounds, cause soiling of commonly used building materials and culturally important items
2 such as statues and works of art. Physical damage from the dry deposition of SO2, particles, and
3 the absorption or adsorption of corrosive agents on deposited particles can also result in the
4 acceleration of the weathering of manmade building and naturally occurring cultural materials.
5
6 • The natural process of metal corrosion from exposure to environmental elements (wind,
7 moisture, sun, temperature fluctuations, etc.) is enhanced by exposure to anthropogenic
8 pollutants, in particular SO2.
9
10 • Dry deposition of SO2 enhances the effects of environmental elements on calcereous stones
11 (limestone, marble, and carbonated cemented) by converting the calcium carbonate (calcite) in
12 the stone to calcium sulphate dihydrate (gypsum). The rate of deterioration is determined by
13 the SO2 concentration, the stone's permeability and moisture content, and the deposition rate.
14
15 • Sulfur dioxide limits the life expectancy of paints by causing discoloration, loss of gloss, and
16 loss of thickness of the paint film layer.
17
18 • A significant detrimental effect of particulate pollution is the soiling of painted surfaces and
19 other building materials. Soiling is a degradation process requiring remediation by cleaning or
20 washing, and depending on the soiled surface, repainting. Soiling decreases the reflectance of a
21 material and reduces the transmission of light through transparent materials. Soiling may
22 reduce the life usefulness of the material soiled.
23
24 E.5.3 Visibility
25 Chapter 4 of this document includes information supplementary to several other significant
26 reviews of the science of visibility, including the 1991 report of the National Acid Precipitation
27 assessment Program, the National Research Council's Protecting Visibility in National Parks
28 and Wilderness Areas (1993), and EPA's 1995 Interim Findings on the Status of Visibility
29 Research. The following points are made in Chapter 4 and/or in the above referenced
30 documents.
31
April 2002 E-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 • The relationships between air quality and visibility are well understood. Ambient fine particles
2 are the major cause of visibility impairment. Significant scientific evidence exists showing that
3 reducing fine particle concentrations will improve visibility.
4
5 • The National Research Council defines visibility qualitatively as "the degree to which the
6 atmosphere is transparent to visible light." This definition may be expressed quantitatively in
7 terms of contrast transmittance. The EPA has defined visibility impairment as a reduction in
8 visual range (the farthest distance at which a large black object can be distinquished against the
9 horizontal sky is the visual range) and/or atmospheric discoloration.
10
11 • Light, as it passes through the atmosphere from a scene to an observer, is both scattered and
12 absorbed. The rate of loss of transmitted light intensity with distance is measured by the light-
13 extinction coefficient, which may be expressed as the sum of the coefficients for: (a) light
14 scattering due to gases; (b) light scattering due to particles; (c) light absorption by gases, and;
15 (d) light absorption by particles. Light scattering by particles is the major component of light
16 extinction. Light absorption by gases is almost entirely due to NO2, and is typically significant
17 only near NO2 sources. Light absorption by particles is primarily caused by elemental carbon.
18
19 • Light scattering efficiency depends on particle size, falling off rapidly for particles below 0.3 or
20 above 1.0 //m in diameter. Therefore, particles in the accumulation mode (of the fine particle
21 mode) are most effective in scattering light and are more important in visibility degradation
22 than either nuclei-mode or coarse-mode particles. Light absorption is not a strong function of
23 particle size. Under exceptional circumstances, such as dust storms, coarse particles can
24 dominate scattering.
25
26 • In addition to reducing the intensity of light carrying information about a scene (transmitted
27 radiance), particles also scatter light into the observer's view. This extraneous light, called air
28 light or path radiance, carries no information about the scene. The competition between these
29 two sources of light, expressed as the ratio of transmitted radiance from the scene to path
30 radiance, determines the contrast transmittance and the visual quality of the view.
31
April 2002 E-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Visibility at any location is affected by air quality and non-air quality related effects. The
2 visibility effects of atmospheric constituents are dependant upon not just the mass of pollutants,
3 but on the size distribution and refractive index of particles, which are strongly influenced by
4 relative humidity. Non-air quality effects include the angle between the sun and the observer's
5 sight path, location of clouds, and reflectivity of the ground. These effects are independent of
6 effects due to changes in atmospheric constituents. Lighting and scene effects can be
7 accounted for by defining a range of these effects when estimating visibility changes due to air
8 quality influences.
9
10 • The relationship between air pollution and the appearance of a scenic view is well understood.
11 Models exist that, given an adequate description of the air quality and non-air quality variables,
12 can produce a simulated photograph that accurately depicts a cloud-free scene as it would
13 appear to a human observer.
14
15 • There are several potential quantitative indicators of visibility, including: (a) fine particle mass
16 and composition (fine particle mass alone provides less of both types of information);
17 (b) scattering by dried ambient particles; (c) scattering by particles under ambient conditions;
18 (d) extinction (calculated from measurements of scattering plus absorption); (e) light extinction
19 measured directly; and (f) contrast transmittance.
20
21 • A new index, the deciview (civ), is now being used as a quantitative measure of haziness. It is
22 related to the light extinction coefficient, bext, by Haziness (dv) = 10 ln(bex/10Mn). The
23 deciview is more nearly linearly related to perceived changes in haze level than either visual
24 range or light extinction. A change of 1 or 2 dv in uniform haze under many viewing conditions
25 will be seen as a small but noticeable change in the appearance of a scene regardless of the
26 initial haze condition.
27
28 • Visibility in the United States is best in the western, intermountain region. Visibility
29 impairment or haziness is greatest in the eastern United States and southern California.
30 Haziness in the eastern United States is caused primarily by atmospheric sulfate. Haziness in
31 southern California is primarily caused by nitrate and organic PM. Nitrates contribute about
April 2002 E-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 40% to the total light extinction in southern California. Nitrates account for 10 to 20% of the
2 total extinction in other areas of the United States.
3
4 • Organics are the second largest contributors to light extinction in most areas in the United
5 States. Organic carbon is the greatest cause of light extinction in the Pacific Northwest,
6 Oregon, Idaho, and Montana, accounting for 40 to 45% of the total extinction. Organic carbon
7 contributes between 15 to 20% to the total extinction in most of the western United States and
8 20 to 30% in the remaining areas of the United States.
9
10 • Light absorption by carbon is relatively insignificant but is highest in the Pacific Northwest
11 (up to 15%) and in the eastern United States (3%).
12
13 • High dust concentrations transported from southern California and the subtropics have
14 contributed to regional haze in the Grand Canyon and other Class I areas in the southwestern
15 United States.
16
17 E.5.4 Global Change Processes and Their Potential Human Health and
18 Environmental Impacts
19 Processes causing global change and their potential environmental and human health
20 impacts have been accorded extensive attention during the past several decades, and they still
21 continue to be of broad national and international concern. This is reflected by extensive
22 research and assessment efforts undertaken since the mid-1970s by U.S. Federal Government
23 Agencies (e.g., NOAA, EPA, CDC, etc.) or via U.S. Federal Interagency programs (e.g., the U.S.
24 Global Change Research Program [USGCRP]). It is also reflected by analogous extensive
25 research and assessment efforts undertaken by numerous other national governments or
26 international collaborative activities, e.g., those coordinated by the Intergovernmental Panel on
27 Climate Change (IPCC), established in the 1980s under the joint auspices of the World
28 Meteorological Organization (WMO) and the United Nations Environment Programme (UNEP).
29 The present discussion of global climate change in Chapter 4 draws upon recent
30 international assessments of (a) processes contributing to stratospheric ozone depletion and the
31 status of progress towards ameliorating the problem (WMO, 1999) and (b) revised qualitative
April 2002 E-3 5 DRAFT-DO NOT QUOTE OR CITE
-------
1 and quantitative projections of likely consequent human health and environmental effects
2 (UNEP, 1998, 2000) — with the findings and conclusions of these assessments being
3 incorporated herein by reference. Also, Chapter 4 presents information on global warming and
4 its potential impacts drawn mainly from extensive assessments contained in the Third
5 Assessment Report of the IPPC (2001) and a series of reports by the USGCRP on projected
6 impacts on the United States or subregions.
7
8 • Atmospheric particles play important roles in two key types of global change phenomena:
9 (1) alterations in the amount of ultraviolet solar radiation (especially UV-B) penetrating
10 through the Earth's atmosphere and reaching its surface, where it can exert a variety of effects
11 on human health, plant and animal biota, and other environmental components; and
12 (2) alterations in the amount of visible solar radiation transmitted through the Earth's
13 atmosphere.
14
15 • Knowledge of factors controlling the transfer of solar radiation in the ultraviolet spectral range
16 is needed for assessing potential biological and environmental impacts associated with
17 exposure to UV-B radiation (290 to 315 nm). Knowledge of the effects of PM on the transfer
18 of radiation in the visible and infrared spectral regions is needed for assessing the relationship
19 between particles and global warming and its environmental and biological impacts.
20
21 PM Effects on Solar Ultraviolet Radiation Transmission Impacts
22 • The main types of deleterious effects hypothesized as likely to result from stratospheric ozone
23 depletion and consequent increased SUVB penetration through the Earth's atmosphere are:
24 (a) Direct Human Health Effects, such as skin damage (sunburn), leading to more rapid aging
25 and increased incidence of skin cancer; ocular effects (retinal damage and increased
26 cataract formation possibly leading to blindness); and suppression of some immune system
27 components (possibly increasing susceptibility to certain infectious diseases or decreasing
28 effectiveness of vaccinations).
29 (b) Agricultural/Ecological Effects, mediated largely through altered biogeochemical cycling
30 resulting in consequent damaging impacts on terrestrial plants (leading to possible reduced
31 yields of rice, other food crops, and commercially important trees, as well as to biodiversity
April 2002 E-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 shifts in natural terrestrial ecosystems); and deleterious effects on aquatic life (including
2 reduced ocean zooplankton and phytoplankton, as important base components of marine
3 food-chains supporting the existence of commercially important, edible fish and other
4 seafood, as well as to other aquatic ecosystem shifts).
5 (c) Indirect Human Health and Ecological Effects, mediated through increased tropospheric
6 ozone formation (and consequent exacerbation of surface-level, ozone-related health and
7 ecological impacts) and alterations in the concentrations of other important trace species,
8 most notably the hydroxyl radical and acidic aerosols.
9 (d) Other Types of Effects, such as faster rates of polymer weathering because of increased
10 UV-B radiation and other effects on man-made commercial materials and cultural artifacts,
11 secondary to climate change or exacerbation of air pollution problems.
12
13 • In contrast to negative impacts projected as likely to be associated with increased UV-B
14 penetration to Earth's surface, some research results are suggestive of possible beneficial
15 effects of increased UV-B radiation. For example, a number of U.S. and international studies
16 report likely protective effects of UV-B radiation with regard to non-skin cancer incidence.
17 These suggests potential relationships between sunlight, vitamin D, and reduced colon cancer
18 and others provide evidence that UV-B radiation may be protective against several types of
19 cancer and some other diseases.
20
21 • From among boundary layer pollutants (e.g., SO2, O3, PM), airborne particles are probably
22 most effective in controlling the amount of SUV-B radiation reaching the Earth's surface. Fine
23 particles are clearly more effective than coarse particles in this regard.
24
25 • Particles scatter and absorb solar radiation in the biologically important UV-B spectral region.
26 The amount of SUV-B reaching the Earth's surface depends in a non-linear way on the content
27 of scattering and absorbing material within airborne particles.
28
29 • Given the above considerations, quantification of projected effects of variations in atmospheric
30 PM on human health or the environment because of the effects of particles on the transmission
31 of solar UV-B radiation requires location-specific evaluations, taking into account composition,
April 2002 E-37 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentration, and internal structure of the particles; temporal variations in atmospheric mixing
2 height and depths of layers containing the particles the abundance of ozone and other absorbers
3 within the planetary boundary layer and the free troposphere. The outcome of such modeling
4 effects would likely vary from location to location in terms of increased or decreased surface
5 level UV-B exposures because of location-specific changes in atmospheric PM concentrations
6 or composition.
7
8 PM Effects on Global Warming Processes and Impacts
9 • During the 20th century global average surface temperatures increased by 0.6 ± 0.2°C. The
10 decade of the 1990s was probably the warmest since 1861. The last few years have been
11 among the warmest on record. The Intergovernmental Panel on Climate Change (IPCC) has
12 attributed most of the warming observed over the past fifty years to human activities. The
13 global average surface temperature is projected to rise by 1.4 to 5.8 °C from 1990 to 2100.
14
15 • There are health effects directly associated with climate change such as increased heat stress
16 and cardiorespiratory failure due to rises in temperature. There are also health effects which
17 are associated with changes in ecosystems and habitats of disease-carrying organisms that are
18 the result of changes in climate variables such as temperature and humidity.
19
20 • Vectorborne diseases such as malaria and dengue fever may extend their ranges in the United
21 States through the northward extension of habitats favorable to their development.
22
23 • Waterborne diseases may likely increase with increasing air and water temperatures, combined
24 with heavy runoff events from agricultural and urban surfaces.
25
26 • The effects of climate change on air quality are also likely to be important, however, these
27 effects are too uncertain to be predicted with any confidence at the present time. Likewise,
28 little is known regarding changes in the effects of air quality on human health under a different
29 climate.
30
April 2002 E-3 8 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Particles, most notably those containing sulfate, exert a direct effect by scattering incoming
2 solar radiation back to space. However, 'black carbon' in particles absorbs solar radiation and
3 as these particles age, their capacity to absorb solar radiation may increase. Some recent
4 estimates indicate that the effect of particles containing 'black carbon' may be twice as large as
5 that given by the latest IPCC report and that the control of 'black carbon' emissions may be an
6 effective means to slow 'greenhouse warming'.
7
8 • Particles also exert an indirect effect on climate by causing an increase in the concentration of
9 cloud droplets which is accompanied by a decrease in their size. The reduction in cloud droplet
10 size tends to increase the liquid water content of the cloud, the lifetime of the cloud and the
11 optical thickness of the cloud. As a result of these processes, clouds reflect more solar
12 radiation back to space. Many of these effects have been observed; for example, cloud droplets
13 in polluted areas tend to be smaller than those formed in clean areas. However, the magnitude
14 of the overall effects on climate, although larger than the direct effects noted above, is more
15 highly uncertain.
16
17 • On a globally averaged basis, radiative cooling due to anthropogenic particles may have
18 substantially offset the radiative heating due to increases in atmospheric concentrations of
19 greenhouse gases such as carbon dioxide, methane, and chlorofluorocarbons.
20
21 • Aerosol lifetimes are also much shorter than the time required for global mixing, therefore,
22 aerosol radiative effects are most likely to exert their influence on a regional rather than on a
23 global basis.
24
25 • The lifetimes of particles in the troposphere are short (days to weeks) compared to the above
26 greenhouse gases (years to over 100 years). Therefore, aerosol concentrations will respond
27 more rapidly to variations in emissions than will the greenhouse gases.
28
29
30
April 2002 E-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 E.6 KEY CONCLUSIONS
2 • Epidemiologic studies show consistent positive associations of exposure to ambient PM with
3 health effects, including mortality and morbidity. The observed associations of ambient PM
4 exposure with health effects must be adjusted for the effects of other environmental or
5 demographic factors, depending on whether the effects are acute or chronic, in order to
6 quantitatively assess the role that may be attributed to PM exposure. Estimates of PM health
7 effects have shown reasonable quantitative consistency in different studies, with only modest
8 sensitivity to different methods of analysis. However, a clearer understanding of specific
9 biologic mechanisms remains to be more fully established.
10
11 • Individuals with cardiovascular or pulmonary disease, especially if they are elderly, are more
12 likely to suffer severe health effects (mortality or hospitalization) related to PM exposure than
13 are healthy young adults. Children and asthmatics are also susceptible to certain PM effects,
14 e.g., increased respiratory symptoms and decreased lung function. Smokers also constitute a
15 population group at increased risk for ambient PM exposure effects.
16
17 • In human populations, daily personal exposures to ambient fine particles are reflected by daily
18 ambient fine particle concentrations measured at a community air-monitoring site. This is
19 consistent with the observed high correlations of personal sulfate exposures with ambient
20 sulfate concentrations. Therefore, community air-monitoring site measurements of fine particle
21 indicators can be useful in PM epidemiology studies. The relationship between personal
22 exposure to thoracic coarse particles and the ambient concentration of thoracic coarse fraction
23 particles is not as strong, making detection of effects due to coarse fraction particles harder to
24 detect in epidemiological studies.
25
26 • Development of a comprehensive biologically-based exposure-dose-response model to aid
27 health risk assessment requires further dosimetry data characterizing differences among species
28 in percent deposition and regional deposition patterns including differences in inhalability,
29 airway geometry, and clearance rates. More information is also required on mechanism(s) of
30 clearance, pathological processes affecting deposition and clearance of particles, and factors
31 which influence the response(s) of respiratory tract tissues to particle burden.
April 2002 E-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 • The percent deposition and regional patterns of deposition depend strongly on particle size.
2 Percent deposition is higher in smaller lungs (children; women), during exercise, and in the
3 functioning parts of the lungs in people with compromised lungs.
4
5 • Estimation of public health impacts of ambient airborne particle exposures in the United States
6 would most credibly require information from exposure-response relationships derived for
7 particular U.S. urban areas, in combination with estimates of exposures to ambient particle
8 concentrations for the general population and/or specific susceptible subgroups (e.g., the
9 elderly) within those particular areas. At the present time, risk assessment is necessarily
10 limited to use of available information from concentration-response relationships relating
11 ambient concentrations to health effects in populations. In view of geographic differences in
12 ambient PM mixtures and demographics, broad generalization and application of some single
13 "best estimate" of relative risk for a given increment in concentration of a given particle
14 indicator (e.g., PM10, PM25, etc.) would be subject to much uncertainty.
15
16 • Toxicology studies of PM using controlled inhalation exposure of humans and laboratory
17 animals, intratracheal instillation in humans and animals, and exposure of human and animal
18 cells grown in culture find numerous biological effects which may be related to adverse health
19 effects. Newer studies are finding different biological effects for a variety of different particle
20 components. Newer studies also are beginning to identify biological mechanisms whereby PM
21 deposited in the lung can produce adverse effects on the cardiovascular and respiratory systems.
22
23 • Epidemiological studies indicate increased health risks associated with exposure to PM, alone
24 or in combination with other air pollutants. PM-related increases in individual health risks are
25 small, but likely significant from an overall public health perspective because of the large
26 numbers of individuals in susceptible risk groups that are exposed to ambient PM.
27 • Numerous new studies, including multicity studies, continue to find a consistent association of
28 PM10 exposure with mortality and various morbidity endpoints, thus substantiating the
29 relationship of PM exposure with various health effects. However, new studies using PM25 as
30 an indicator find higher statistical significance and higher excess risk for PM2 5 compared to
April 2002 E-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 PM10. Several studies have also observed statistically associations of PM10_25 with health
2 effects.
3
4 • Epidemiologic studies, in which factors identified with source categories or individual
5 chemical components of PM have been used as indicators, also show significant associations
6 with nitrate, sulfate, various indicators of elemental carbon, the organic component of PM, and
7 some elements. Source category factors, attributed to PM associated with regional sulfate,
8 vehicular traffic, vegetative combustion, and oil burning have also been found to be
9 significantly associated with mortality.
10
11 • Data from multicity studies, comparisons of effects found in single and multiple regressions,
12 exposure relationships, and toxicity suggest that the gaseous co-pollutants (CO, NO3, SO2,O3)
13 are not responsible for the relationships observed with PM indicators in community, time-series
14 epidemiology. This does not indicate lack of an independent association of gaseous
15 co-pollutants with health effects, but rather suggests that they may be surrogates for PM (due to
16 significant correlations with PM) rather than confounders.
17
18 • Fine and thoracic coarse PM, indexed respectively by PM25 and PM10_25, should be considered
19 as separate subclasses of PM. Considerations of emissions sources, atmospheric chemistry,
20 physical behavior, exposure relationships, respiratory deposition, toxicologic findings, and
21 epidemiologic observations argue for monitoring fine and thoracic coarse particles separately.
22
23 • Assessment of health risk in epidemiologic studies of ambient air pollutants, including PM, has
24 relied largely on studies that focus on changes in health risks that occur in relation to normal
25 changes in ambient air pollutant concentrations. Further evidence of the effects of air pollution
26 on health may be deduced from intervention studies, i.e, studies of changes in health effects
27 that occur when air pollution concentrations have been temporarily or permanently reduced
28 through regulatory action, industrial shutdown, or other intervening factor(s). Only a few
29 epidemiologic intervention studies are available, however, taken together, these studies lend
30 confidence that further reduction of ambient air pollution exposures in the U. S. would benefit
31 public health. It is likely that such reduction would bring about both respiratory and
April 2002 E-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 cardiovascular heath benefits. Available studies also give reason to expect that further
2 reductions in both particulate and gaseous air pollutants would benefit health. On balance,
3 these studies suggest that selective reduction in ambient PM concentrations might well bring
4 about greater benefit than would selective reduction in concentrations of other ambient criteria
5 air pollutants. Furthermore, the experimental studies of Utah Valley filter extracts points to
6 PM-associated metals as a likely cause or promoter of at least some of the health disorders
7 associated with ambient PM.
8
9 • The weight of evidence, from exposure, dosimetry, toxicology, and epidemiology, leans toward
10 the conclusion that PM, especially fine PM, is the primary contributor to a variety of adverse
11 health effects associated with air pollution. However, there are difficult technical issues in
12 separating the effects of fine and coarse particles and in separating particle effects from
13 possible effects of gaseous co-pollutants.
14
April 2002 E-43 DRAFT-DO NOT QUOTE OR CITE
-------
i 1. INTRODUCTION
2
3
4 This document is an update of "Air Quality Criteria for Particulate Matter" published by the
5 U.S. Environmental Protection Agency (EPA) in 1996, and it will serve as the basis for
6 Congress!onally-mandated periodic review of the National Ambient Air Quality Standards
7 (NAAQS) for particulate matter (PM). The present document critically assesses the latest
8 scientific information relative to determining the health and welfare effects associated with
9 exposure to various concentrations of PM in ambient air. The document builds upon the
10 assessment in the previous 1996 EPA Air Quality Criteria Document for Particulate Matter by
11 focusing on assessment and integration of information most relevant to PM NAAQS criteria
12 development, based on pertinent literature mainly available through December 2001, as well as
13 some newly emerging studies published in early 2002. This introductory chapter presents a brief
14 summary of legislative requirements and history of the PM NAAQS, provides an overview of
15 issues addressed and procedures utilized in the preparation of the present document, and provides
16 orientation to the general organizational structure of this document.
17
18
19 1.1 LEGISLATIVE REQUIREMENTS
20 As indicated in U.S. Code (1991), the U.S. Clean Air Act (CAA), Sections 108 and 109
21 (42 U.S.C. Sections 7408 and 7409) govern the establishment, review, and revision of National
22 Ambient Air Quality Standards (NAAQS). Section 108a directs the EPA Administrator to list
23 pollutants, which, in the Administrator's judgement, cause or contribute to air pollution which
24 may reasonably be anticipated to endanger either public health or welfare and which are emitted
25 by numerous (or diverse) stationary or mobile sources. The air quality criteria are to reflect the
26 latest scientific information useful in indicating the kind and extent of exposure-related effects on
27 public health and welfare that may be expected from the presence of the pollutant in ambient air.
28 Section 109(a) and (b) directs the Administrator of EPA to propose and promulgate
29 "primary" and "secondary" NAAQS for pollutants identified under Section 108. Section
30 109(b)(l) defines a primary standard as a level of air quality, the attainment and maintenance of
April 2002 1-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 which, in the judgement of the Administrator, based on the criteria and allowing for an adequate
2 margin of safety, is requisite to protect the public health. A secondary standard, in contrast, is
3 one which, again in the Administrator's judgement, is requisite to protect public welfare from
4 any known or anticipated adverse effects associated with the presence of such pollutants.
5 Welfare effects include effects on vegetation, crops, soils, water, animals, manufactured
6 materials, visibility, weather, and climate, as well as damage to and deterioration of property,
7 hazards to transportation, and effects on economic value and personal comfort and well-being
8 (as per CAA Section 302h). Section 109(d) also requires periodic review and, as appropriate,
9 revision of existing criteria and standards; and it requires an independent committee of non-EPA
10 experts, the Clean Air Scientific Advisory Committee (CASAC), to provide advice and
11 recommendations to the EPA Administrator regarding the scientific soundness and
12 appropriateness of criteria and NAAQS for PM and other "criteria air pollutants" (e.g., ozone,
13 nitrogen oxides, sulfur oxides, carbon monoxide, lead) regulated under CAA Sections 108-109.
14
15
16 1.2 HISTORY OF PREVIOUS PM CRITERIA AND NAAQS REVIEWS
17 "Particulate matter" is the generic term for a broad class of physically and chemically
18 diverse substances that exist in ambient air as discrete particles (liquid droplets or solids) over a
19 wide range of sizes. These airborne particles originate from a variety of stationary and mobile
20 sources. Primary particles are emitted directly into ambient air; whereas secondary particles are
21 formed in the atmosphere by transformation of gaseous emissions such as sulfur oxides (SOX),
22 nitrogen oxides (NOX), and volatile organic compounds (VOCs). The physical and chemical
23 properties of PM vary greatly with time, region, meteorology, and source category, thus
24 complicating assessment of ambient PM health and welfare effects. Particles in ambient air are
25 usually distributed bimodally in two somewhat overlapping size categories: (1) fine (diameter
26 generally less than 2.5 //m) and (2) coarse (diameter generally greater than 2.5 //m). Particles in
27 these two size fractions tend to differ in terms of sources of origin, composition, and behavior in
28 the atmosphere and human respiratory tract.
29 EPA first promulgated primary and secondary NAAQS for PM on April 30, 1971 (Federal
30 Register, 1971). These standards measured PM as "total suspended particulate" (TSP), which
31 refers to ambient PM up to a nominal size of 25 to 45 micrometers (//m). The primary standards
April 2002 1-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 for PM (measured as TSP) were 260 //g/m3 (24-h average), not to be exceeded more than once
2 per year, and 75 //g/m3 (annual geometric mean). The secondary standard (measured as TSP)
3 was 150 //g/m3 (24-h average), not to be exceeded more than once per year.
4 EPA completed the next review of PM air quality criteria and standards in July 1987,
5 revising the 1971 standards to protect against adverse health effects of inhalable airborne
6 particles with an upper 50% cut-point of 10-//m aerodynamic diameter (PM10), which can be
7 deposited in the lower (thoracic) regions of the human respiratory tract (Federal Register, 1987).
8 EPA established identical primary and secondary PM10 standards for two averaging times:
9 150 //g/m3 (24-h average), with no more than one expected exceedance per year and 50 //g/m3
10 (expected annual arithmetic mean), averaged over three years.
11
12 1.2.1 The 1997 PM NAAQS Revision
13 The last previous review of the air quality criteria and standards for PM was initiated in
14 April 1994 by EPA announcing its intention to develop revised Air Quality Criteria for
15 Paniculate Matter (henceforth, the "PM Air Quality Criteria Document" or PM AQCD). Several
16 workshops were held by EPA's Environmental Criteria and Assessment Office in Research
17 Triangle Park, NC (ECAO-RTP) in November 1994 and January 1995 to discuss important new
18 health effects information useful in preparing initial PM AQCD draft materials. Also, plans for
19 review of the PM criteria and standards under a highly accelerated, court-ordered schedule were
20 presented by EPA at a public meeting of the CAS AC in December 1994. A court order entered
21 in American Lung Association v. Browner, CIV-93-643-TUC-ACM (U.S. District Court of
22 Arizona, 1995), as subsequently modified, required publication of EPA's final decision on the
23 review of the PM NAAQS by July 19, 1997.
24 Several external review drafts of the revised PM AQCD were prepared by the RTF
25 Division of EPA's newly created National Center for Environmental Assessment (i.e., by
26 NCEA-RTP, the successor office to ECAO-RTP), and each were made available for public
27 comment followed by CAS AC review (at public meetings held in August 1995, December 1995,
28 and February 1996). The CAS AC came to closure on its review of the PM AQCD in early 1996,
29 advising the EPA Administrator in a March 15, 1996 closure letter (Wolff, 1996) that "although
30 our understanding of the health effects of PM is far from complete, a revised Criteria Document
31 which incorporates the Panel's latest comments will provide an adequate review of the available
April 2002 1-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 scientific data and relevant studies of PM." Revisions made in response to public and CAS AC
2 comments were then incorporated by NCEA-RTP, as appropriate, into the final 1996 PM AQCD
3 (U.S. Environmental Protection Agency, 1996a). A PM Staff Paper (SP), prepared by the Office
4 of Air Quality Planning and Standards (OAQPS) within EPA's Office of Air and Radiation
5 (OAR), drew upon the 1996 PM AQCD and other exposure and risk assessments to pose options
6 for PM NAAQS decisions. The PM SP also underwent public comment and CASAC review,
7 with consequent revision to its July 1996 final form (U.S. Environmental Protection Agency,
8 1996b). The SP analyses served as key inputs to subsequently published proposals (Federal
9 Register, 1996) for revision of the primary PM NAAQS.
10 Taking into account information and assessments presented in the 1996 PM AQCD and
11 PM SP, advice and recommendations of CASAC, and public comments received on the proposed
12 revisions, the EPA Administrator revised the PM NAAQS by adding new PM2 5 standards and by
13 revising the form of the 24-h PM10 standard (Federal Register, 1997a). Specifically, on July 18,
14 1997, the Administrator promulgated the following revisions to the PM NAAQS:
15 (1) The suite of PM standards was revised to include an annual primary PM25 standard and a
16 24-h PM2 5 standard.
17 (2) The 24-h PM25 standard is met when the 3-year average of the 98th percentile of 24-h PM2 5
18 concentrations at each population-oriented monitor within an area is less than or equal to
19 65 //g/m3, with fractional parts of 0.5 or greater rounding up.
20 (3) The annual PM2 5 standard is met when the 3-year average of the annual arithmetic mean
21 PM25 concentrations, from single or multiple community-oriented monitors, is less than or
22 equal to 15 //g/m3, with fractional parts of 0.05 or greater rounding up.
23 (4) The form of the 24-h PM10 (150 //g/m3) standard was revised to be based on the 3-year
24 average of the 99th percentile of 24-h PM10 concentrations at each monitor within an area.
25 (5) In addition, the Administrator retained the annual PM10 standard at the level of 50 //g/m3,
26 which is met when the 3-year average of the annual arithmetic mean PM10 concentrations at
27 each monitor within an area is less than or equal to 50 //g/m3, with fractional parts of 0.5 or
28 greater rounding up.
29 The principal focus of the last review of the air quality criteria and standards for PM was on
30 recent epidemiological evidence reporting associations between ambient concentrations of PM
31 and a range of serious health effects. Special attention was given to several size-specific classes
April 2002 1-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 of particles, including PM10 and the principal fractions of PM10, referred to as the fine (PM25) and
2 coarse (PM10_2 5) fractions. PM2 5 refers to particles with an upper 50% cutpoint of 2.5 fj.m
3 aerodynamic diameter. PM10_25 refers to those particles with an upper 50% cutpoint of 10 //m
4 and a lower 50% cut point of 2.5 //m aerodynamic diameter. In other words, the coarse fraction
5 (PM10_25) refers to the inhalable particles that remain if fine (PM25) particles are removed from a
6 sample of PM10 particles. As discussed in the 1996 PM AQCD, fine and coarse fraction particles
7 can be differentiated by their sources and formation processes and by their chemical and physical
8 properties, including behavior in the atmosphere. Detailed discussions of atmospheric formation,
9 ambient concentrations, and health effects of ambient air PM, as well as quantitative estimates of
10 human health risks associated with exposure to ambient air PM, can be found in the 1996 PM
11 AQCD and the 1996 OAQPS SP (U.S. Environmental Protection Agency, 1996a,b).
12
13 1.2.2 Coordinated Particulate Matter Research Program
14 Shortly after promulgation of the 1997 PM NAAQS decisions, NCEA-RTP published a PM
15 Health Risk Research Needs Document (U.S. Environmental Protection Agency, 1998a) that
16 identified research needed to improve scientific information supporting future reviews of the PM
17 NAAQS. The document aimed to provide a foundation for PM research coordination among
18 Federal agencies and other research organizations and provided input to later National Research
19 Council (NRC) PM research deliberations. The Office of Research and Development (ORD) of
20 EPA also moved quickly to broaden its ongoing PM research activities by developing, in
21 partnership with other Federal agencies, a coordinated interagency PM research program. This
22 interagency program has and continues to focus mainly on expanding scientific knowledge of
23 ambient PM exposure and health effects, as well as including development of improved
24 monitoring methods and cost-effective mitigation strategies. The interagency effort also
25 promotes substantially expanded coordination with other research organizations, including the
26 Health Effects Institute (HEI) and other state-, university-, and industry-sponsored research
27 groups. Beginning in the fall of 1997, public participation was and continues to be encouraged
28 through workshops and review of program documentation.
29 In response to Congressional requirements in EPA's Fiscal Year 1998 Appropriation, the
30 NRC established its Committee on Research Priorities for Airborne Particulate Matter in January
31 1998. This NRC PM Research Committee's charge is to identify the most important research
April 2002 1-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 priorities relevant to setting particulate matter standards, to develop a conceptual plan for PM
2 research, and to monitor research progress toward improved understanding of the relationship
3 between PM and public health. The Committee issued its first report in early 1998 (National
4 Research Council, 1998), a second one in 1999 (National Research Council, 1999), and a third
5 one in 2001 (National Research Council, 2001).
6 In the above-noted series of reports, the NRC PM Research Committee recommended that
7 expanded PM research efforts be planned and carried out in relation to a general conceptual
8 framework as shown in Figure 1-1. That framework essentially calls for research aimed at:
9 (a) identifying sources of airborne particles or gaseous precursor emissions and characterization
10 of processes involved in atmospheric transformation, transport, and fate of ambient PM;
11 (b) delineation of temporal and spatial patterns of air quality indicators (e.g., PM25, PM10_25,
12 PM10 mass concentrations) of ambient PM and apportionment of observed variations in such
13 ambient PM indicators to various emission sources; (c) characterization of human exposures to
14 ambient PM as one important component of total personal exposure to particles, as modified by
15 time-activity patterns and varying microenvironmental exposure to particles of indoor or ambient
16 origin; (d) characterization of resulting respiratory tract deposition, clearance, retention, and
17 disposition of inhaled particles, as determinants of dose to target tissues (e.g., locally in the lungs
18 or via systemic translocation to the heart or other organs); (e) delineation of mechanisms of
19 damage and repair plausibly leading to (f) human health responses, as extrapolated from or
20 quantified by experimental animal or human exposure (toxicology) studies and/or observational
21 (epidemiology) studies.
22 Research conducted under a PM Research Program structured in relation to the conceptual
23 framework shown in Figure 1-1 would be expected (a) to reduce key scientific uncertainties
24 regarding interrelationships between PM sources, ambient concentrations, exposures, dose to
25 target tissues, and resulting health effects and (b) thereby improve the scientific underpinnings
26 for both current and future periodic PM criteria/NAAQS reviews. Table 1-1 highlights some
27 types of key uncertainties identified by the NRC PM Research Committee in relation to elements
28 of the source-to-response conceptual framework illustrated in Figure 1-1. The NRC Committee
29 went on to delineate a series of 10 research topics that they recommended be addressed in an
30 expanded PM research program aimed at answering a set of broadly stated questions, as shown in
31 Table 1-2.
April 2002 1-6 DRAFT-DO NOT QUOTE OR CITE
-------
Sources of Airborne
Particulate Matter
Or Gaseous
Precursor Emissions
Indicator in
Ambient (Outdoor) Air
(e.g. Mass
Concentration)
Mechanisms determining emissions,
chemical transformation (including
formation of secondary particles from
gaseous precursors), and
transport in air
Human time-activity
patterns, indoor (or
microenvironmental)
sources and sinks of
particulate matter
Deposition,
clearance, retention
and disposition of
particulate matter
presented to an
individual
Mechanisms of
damage and repair
Figure 1-1. A general framework for integrating particulate-matter research. Note that
this figure is not intended to represent a framework for research management.
Such a framework would include multiple pathways for the flow of
information.
Source: National Research Council (2001), as modified from NRC (1983, 1994), Lioy (1990), and Sexton et al.
(1992).
1
2
3
4
5
9
10
11
12
13
14
The EPA's PM Research Program is structured to address the topics shown in Table 1-2;
and it includes, for example, studies to improve understanding of the formation and composition
of fine PM, improved measurements and estimation of population exposures to ambient PM, the
characteristics or components of PM that are responsible for its health effects, and the
mechanisms by which these effects are produced. Specific EPA research efforts include
controlled human exposure studies, in vivo and in vitro toxicology, epidemiology, atmospheric
sciences including monitoring and modeling studies, development of data on emissions of fine
particles from stationary and mobile sources, and identification and evaluation of risk
management options. The results from these efforts, as well as related efforts by other Federal
agencies and the general scientific community during the past several years, have substantially
enhanced the scientific and technical bases for future decisions on the PM NAAQS, as well as for
implementation of PM monitoring and control efforts that are beyond the scope of this document.
April 2002
1-7
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 1-1. KEY SCIENTIFIC UNCERTAINTIES RELATED TO THE
SOURCE-TO-RESPONSE FRAMEWORK
Source ^ Concentration (or other indicator)
• Contribution of various emission sources to ambient and indoor paniculate matter concentrations
• Relative contribution of various sources to the most toxic components of paniculate matter
Concentration (indicator) ^ Exposure
• Relationship between ambient (indoor) paniculate matter and the composition of particles to which people
are exposed
• Contribution of ambient paniculate matter to total personal exposure for:
- Susceptible subpopulations
- General population
• Variation in relationship of ambient paniculate matter concentrations to human exposure by place
• Variation in contribution of ambient paniculate matter concentrations to total human exposure over time
• Covariance of paniculate matter exposures with exposures to other pollutants
• Relationships between outdoor ambient and personal exposures for paniculate matter and copollutants
Exposure ^ Dose
• Relationship between inhaled concentration and dose of paniculate matter and constituents at the tissue
level in susceptible subjects
-Asthma
- Chronic obstructive pulmonary disease (COPD)
- Heart disease
- Age: infants and elderly
- Others
Dose ^ Response
• Mechanisms linking morbidity and mortality to paniculate matter dose to or via the lungs
- Inflammation
- Host defenses
- Neural mechanisms
Source: National Research Council (2001).
April 2002 1-8 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 1-2. RESEARCH TOPICS AND QUESTIONS RECOMMENDED BY
NATIONAL RESEARCH COUNCIL (NRC) TO BE ADDRESSED BY EXPANDED
PM RESEARCH PROGRAM
RESEARCH TOPIC 1. OUTDOOR MEASURES VERSUS ACTUAL HUMAN EXPOSURES
• What are the quantitative relationships between concentrations of particulate matter and gaseous
copollutants measured at stationary outdoor air monitoring sites and the contributions of these
concentrations to actual personal exposures, especially for subpopulations and individuals?
RESEARCH TOPIC 2. EXPOSURES OF SUSCEPTIBLE SUBPOPULATIONS TO TOXIC
PARTICULATE MATTER COMPONENTS
• What are the exposures to biologically important constituents and specific characteristics of particulate
matter that cause responses in potentially susceptible subpopulations and the general population?
RESEARCH TOPIC 3. CHARACTERIZATION OF EMISSION SOURCES
• What are the size distribution, chemical composition, and mass-emission rates of particulate matter
emitted from the collection of primary-particle sources in the United States, and what are the emissions
of reactive gases that lead to secondary particle formation through atmospheric chemical reactions?
RESEARCH TOPIC 4. AIR-QUALITY MODEL DEVELOPMENT AND TESTING
• What are the linkages between emission sources and ambient concentrations of the biologically important
components of particulate matter?
RESEARCH TOPIC 5. ASSESSMENT OF HAZARDOUS PARTICULATE MATTER COMPONENTS
• What is the role ofphysicochemical characteristics of particulate matter in eliciting adverse health effects?
RESEARCH TOPIC 6. DOSIMETRY: DEPOSITION AND FATE OF PARTICLES IN THE
RESPIRATORY TRACT
• What are the deposition patterns and fate of particles in the respiratory tract of individuals belonging to
presumed susceptible subpopulations?
RESEARCH TOPIC 7. COMBINED EFFECTS OF PARTICULATE MATTER AND GASEOUS
POLLUTANTS
• How can the effects of particulate matter be disentangled from the effects of other pollutants? How can the
effects of long-term exposure to particulate matter and other pollutants be better understood?
RESEARCH TOPIC 8. SUSCEPTIBLE SUBPOPULATIONS
• What subpopulations are at increased risk of adverse health outcomes from particulate matter?
RESEARCH TOPIC 9. MECHANISMS OF INJURY
• What are the underlying mechanisms (local pulmonary and systemic) that can explain the epidemiological
findings of mortality/morbidity associated with exposure to ambient particulate matter?
RESEARCH TOPIC 10. ANALYSIS AND MEASUREMENT
• To what extent does the choice of statistical methods in the analysis of data from epidemiological studies
influence estimates of health risks from exposures to particulate matter? Can existing methods be
improved? What is the effect of measurement error and misclassification on estimates of the association
between air pollution and health?
Source: National Research Council (2001).
April 2002 1-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 1.3 CURRENT PM CRITERIA AND NAAQS REVIEW
2 1.3.1 Key Milestones
3 As with other NAAQS reviews, a rigorous assessment of relevant scientific information is
4 to be presented in this updated, revised PM AQCD. As shown in Table 1-3, development of the
5 document has involved substantial external peer review through (a) public workshops involving
6 the general aerosol scientific community, (b) iterative reviews of successive drafts by CASAC,
7 and (c) comments from the public. The final document will reflect input received through these
8 reviews and will serve to evaluate and integrate the latest available scientific information to
9 ensure that the review of the PM standards is based on rigorous evaluation of the available
10 science. An earlier (October 1999) First External Review Draft of this updated document was
11 released in the fall of 1999 for public comment and CASAC review. A Second External Review
12 Draft (March 2001) took into account the earlier public comments and the December 1999
13 CASAC review. This Third External Review Draft similarly takes into account prior public
14 comments and CASAC recommendations from its July 2001 review, and it includes
15 consideration of relevant new peer-reviewed scientific studies published or accepted for
16 publication mainly through December 2001, as well as some newly emerging key studies
17 published in early 2002. Following a 60-day public comment period, it is to be reviewed by
18 CASAC at a public meeting in July 2002.
19 After CASAC review of the First External Review Draft of this revised PM AQCD in
20 December 1999, EPA's OAQPS started to prepare the associated PM Staff Paper. A preliminary
21 draft SP was made available to the public and CASAC for review at their July 2001 meeting.
22 The next draft PM SP will, to the extent possible, draw on the updated findings and conclusions
23 from this Third Draft of the PM AQCD and will also undergo further public comment and
24 CASAC review (scheduled for September 2002). Ultimately drawing on information in the final
25 version of this newly revised PM AQCD, the PM SP will evaluate policy implications of the key
26 studies and scientific findings contained in the AQCD, present related staff analyses of air quality
27 and human health risk, and identify critical elements that EPA staff believes should be
28 considered in reviewing the PM standards. The PM SP is intended to bridge the gap between the
29 scientific review in the AQCD and the public health and welfare policy judgements required of
30
April 2002 1-10 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 1-3. SCHEDULE FOR DEVELOPMENT OF THE CURRENT REVISED
PARTICULATE MATTER AIR QUALITY CRITERIA DOCUMENT (PM AQCD)
Major Milestones
Target Dates
PM NAAQS Review Plan to CASAC
Prepare AQCD Development Plan
Begin Literature Search
Federal Register Call for Information/Sources Sought
CASAC Meeting on AQCD Development Plan
Prepare Workshop Drafts of Chapters
Peer Review Workshop
Prepare External Review Draft AQCD
First External Review Draft
Public Comment Period on Draft AQCD
CASAC Meeting on First Draft AQCD
Second External Review Draft
Public Comment Period on Second Draft
CASAC Meeting on Second Draft
Third External Review Draft
Public Comment Period on Third Draft
CASAC Meeting on Third Draft
Final PM AQCD
October 1997
November 1997 to January 1998
February 1998
April 1998
May 1998
May to December 1998
April 1999
March to September 1999
October 1999
October 1999 to January 2000
December 1999
March 2001
April to July 2001
July 2001
April 2002
May to July 2002
July 2002
Oct/Nov 2002
1 the Administrator in reviewing the PM NAAQS. In doing so, the PM SP will include staff
2 conclusions and recommendations of options for the Administrator's consideration.
3 Based on the final versions of the PM AQCD and the PM SP and on the advice of CASAC,
4 the Administrator will propose to retain or revise the current PM NAAQS and provide
5 opportunities for public comment and CASAC review of the proposed decisions. Taking into
6 account public comments and CASAC advice, the Administrator will then make final decisions
7 on the current PM NAAQS, which are now expected to be issued by December 31, 2003.
April 2002
1-11
DRAFT-DO NOT QUOTE OR CITE
-------
1 1.3.2 Methods and Procedures for Document Preparation
2 The procedures for developing this revised PM AQCD build on the experience and
3 methods derived from the most recent previous PM, Ozone, and CO AQCD preparation efforts.
4 Briefly, the respective responsibilities for production of the present PM AQCD are as follows.
5 An NCEA-RTP PM team was formed and is responsible for developing and implementing the
6 project plan for preparation of the PM AQCD, taking into account inputs from individuals in
7 other EPA program and policy offices identified as part of the EPA PM Work Group. The
8 resulting project plan (i.e., the PM Document Development Plan) was then discussed with
9 CAS AC (May 1998) and appropriately revised. An ongoing literature search has continued to be
10 conducted to identify, to the extent possible, all pertinent PM literature published since early
11 1996. Additionally, EPA published (1) a request for information in the Federal Register asking
12 for recently available research information on PM that may not yet be published and (2) a request
13 for individuals with the appropriate type and level of expertise to contribute to the writing of PM
14 AQCD materials to identify themselves (U.S. Environmental Protection Agency, 1998b).
15 Specific authors of chapters or sections of the proposed document were selected on the basis of
16 their expertise on the subject areas and their familiarity with the relevant literature; these include
17 both EPA and non-EPA scientific experts. The project team defined critical issues and topics to
18 be addressed by the authors and provided direction in order to emphasize evaluation of those
19 studies most clearly identified as important for standard setting. It should be noted that materials
20 contributed by non-EPA authors are incorporated and, at times, modified by EPA PM team staff
21 to reflect internal and/or external review comments, e.g., by the public or CASAC, and that EPA
22 is responsible for the ultimate content of the PM AQCD.
23 The main focus of this revised criteria document is the evaluation and interpretation of
24 pertinent atmospheric science information, air quality data, human exposure information, and
25 health and welfare effects information newly published since that assessed in the 1996 PM
26 AQCD and likely to be useful in deriving criteria for PM NAAQS. Initial draft versions of
27 AQCD chapters were evaluated via expert workshops and/or expert written peer reviews that
28 focused on the selection of pertinent studies included in the chapters, the potential need for
29 additional information to be added to the chapters, and the quality of the summarization and
30 interpretation of the literature. The authors of the draft chapters then revised them on the basis of
31 the workshop and/or written expert review recommendations. These and other integrative
April 2002 1-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 summary materials were incorporated into the First External Review Draft of the PM AQCD
2 (October 1999), which was released for public comment and reviewed at a December 1999
3 CASAC public meeting.
4 In order to foster timely presentation and publication of newly emerging PM research
5 findings, EPA co-sponsored an Air and Waste Management Association International Speciality
6 Conference, entitled "PM 2000: Paniculate Matter and Health", which was held in January 2000
7 in Charleston, SC. The conference was co-sponsored in cooperation with several other
8 government agencies and/or private organizations that also fund PM research. Topics covered
9 included new research results concerning the latest advances in PM atmospheric sciences (e.g.,
10 PM formation, transport, transformation), PM exposure, PM dosimetry and extrapolation
11 modeling, PM toxicology (e.g., mechanisms, laboratory animal models, human clinical
12 responses), and PM epidemiology. The main purpose of the conference was to facilitate having
13 the latest scientific information available in time for incorporation as quickly as possible into the
14 Second External Review Draft of this revised PM AQCD. Hence, arrangements were made for
15 scientists to submit written manuscripts on papers or posters presented at the PM 2000
16 Conference for expedited peer-review by several major journals, so that decisions on acceptance
17 for publication could be made by mid-2000. The evaluations and findings set forth in the Second
18 External Review Draft (March 2001) of the revised PM AQCD included consideration of such
19 published PM 2000 papers and extensive additional information published elsewhere since the
20 previous First External Review Draft; it also reflected public and CASAC comments on that First
21 Draft. The Second External Review Draft (March 2001) was then reviewed by CASAC in July
22 2001. Further revisions incorporated into this Third External Review Draft (April 2002) reflect
23 both public comment and CASAC review of the Second Draft, as well as assessment of
24 additional pertinent information published since that addressed in the Second Draft. The final
25 version of the newly revised PM AQCD will then incorporate changes made in response to
26 public comments and CASAC review of this Third External Review Draft.
27
28 1.3.3 Approach
29 The approach to development of this revised PM AQCD is somewhat different from that
30 used for previous criteria documents. Because the most recent prior document the 1996 PM
31 AQCD (U.S. Environmental Protection Agency, 1996a) provides an extensive discussion of most
April 2002 1-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 topic areas, this new document focuses more specifically on critical issues that have been
2 identified by the NRC as needing more research in order to improve the scientific bases (criteria)
3 for PM NAAQS (see Section 1.2.2. above).
4 An initial step was to focus on selection of pertinent issues to include in the document as
5 the basis for the development of PM NAAQS criteria. Preliminary issues were identified by the
6 NCEA PM Team via input from other EPA program and policy offices, as derived from the 1996
7 PM AQCD and SP, their CAS AC and public reviews, from the 1997 standard promulgation
8 process, and from the 1998 PM Research Needs Document (alluded to in Section 1.2.2). Still,
9 further identification and refinement of issues resulted from NRC review and reports on PM
10 research priorities (also discussed in Section 1.2.2 above). The CASAC review of the PM
11 AQCD Development Plan and public comments on earlier draft AQCD materials at various
12 stages of their development have also played important roles in issue identification.
13 In developing draft materials for inclusion in the revised PM AQCD, detailed review of key
14 new research was undertaken as a first step. However, instead of presenting a comprehensive
15 review of all the literature, emphasis in this revised AQCD is placed on (1) first, the concise
16 summary of key findings derived from previous PM criteria reviews and, then, (2) summarization
17 and evaluation of the most pertinent new key information, with greater emphasis on more
18 interpretive assessment—an approach reflecting CASAC recommendations. To aid in the
19 development of a concise document, compilation of summary tables of relevant new literature
20 published since completion of the previous 1996 PM AQCD and selective text discussion of that
21 literature has been undertaken, with increased emphasis being placed in text discussions on
22 interpretive evaluation and integration of key points derived from the newly summarized research
23 results.
24
25 1.3.4 Key Human Health Issues of Concern
26 The present document reviews and assesses available data bearing on each of the broad
27 topics or issues identified below:
28 (1) Causality. Evaluation of the evidence for or against a causal relationship between health
29 outcomes and ambient PM and/or specific PM physical-chemical components.
April 2002 1-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Specific components of interest include: (a) size classes, such as PM10, PM10_25, PM25,
2 and ultrafme particles, and (b) chemical components, such as transition metals, acidity,
3 sulfates, nitrates, and organics.
4 • Expand review of foundations of causal inference for associated PM air pollution health
5 effects.
6 • Assess new long-term PM exposure and health data to broaden interpretation of long-term
7 exposure findings.
8 • Review data exploring potential mechanisms of response to PM physical-chemical
9 characteristics, response pathways, and exposure-dose-response relationships (laboratory
10 and clinical research).
11 (2) Uncertainties. In carrying out overall assessment, address the following types of
12 uncertainty.
13 • Uncertainties between stationary PM monitoring instruments and personal exposure to PM
14 of ambient origin, especially for susceptible groups and their related activity patterns.
15 Specific topics include measurement error in outdoor monitors themselves, use of central
16 monitors for estimates of community concentrations, and the use of community
17 concentrations as a surrogate for personal exposure to particles of ambient origin.
18 • Uncertainties related to particulate matter size fraction, particle number, surface area, and
19 content of semivolatile components.
20 • Uncertainties about the effects of long-term PM exposure, such as life shortening, and
21 development and progression of disease.
22 • Uncertainties because of coexposure to other pollutants such as O3, SO2, CO, and NO2,
23 and because of meterological factors.
24 • Uncertainties because of potential confounding in epidemiologic studies (e.g., economic
25 factors, demographic and lifestyle attributes, genetic susceptibility factors, occupational
26 exposure, medical care).
27 • Uncertainty about shape of concentration-response (CR) relationships and associated
28 community risks (linear and threshold models for CR).
29 • Uncertainty about methods for synthesis of health outcome studies and evaluation of
30 sensitivity and confounding aspects, including but not limited to meta-analyses.
April 2002 1-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 (3) Biological Mechanisms of Action. Evaluate data examining mechanisms underlying health
2 outcomes of PM. Mechanistic information aids judgment about causality.
3 • New studies have examined mechanisms of action of PM constituents, including transition
4 metals, airborne allergens, and the generation of reactive oxygen species. Different cell
5 types have differing responses to PM components.
6 • Newly published studies also have identified potential mechanisms for the production of
7 cardiac arrhythmias by PM constituents, especially in animal models of disease and
8 suggest that particular attention should be accorded to PM metal constituents.
9 • Although many new animal toxicology studies involve instillation in the lung of
10 previously collected particles and this technique is appropriate to study mechanisms of
11 action, extrapolation to human equivalent exposure/doses is uncertain.
12 • Ongoing work on the effects of lung inflammation and PM phagocytosis on subsequent
13 systemic effects, especially cardiac or vascular effects, is needed to provide further
14 information on the relationship between inhaled pollutants and cardiac events.
15 • Interpretation of concentrated ambient particles (CAPs) studies. Newly available
16 information is examined from toxicology studies using devices that concentrate
17 (to variable extents) ambient PM to determine PM concentration-response relationships.
18 Again, difficulties exist with regard to quantitative extrapolation to comparable human
19 exposures to ambient PM.
20 (4) Susceptible Populations. Examine health outcome data to determine specific risk groups
21 that are more susceptible than normal healthy adults to adverse effects from PM exposure.
22 • Preexisting respiratory or cardiovascular disease in conjunction with advanced age appear
23 to be important factors in PM mortality susceptibility.
24 • For morbidity health endpoints, children and asthmatics potentially may display increased
25 sensitivity to PM exposure. Data will be examined for coherence.
26 • Patterns of respiratory tract deposition, clearance, and retention in susceptible populations
27 have been studied recently and provide evidence of differences in respiratory tract PM
28 deposition for children and small-sized adults and for those with lung diseases.
29 • Animal models of lung disease exposed to PM constituents suggest a role for PM in
30 cardiac death.
31
April 2002 1-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 (5) Environmental Effects. Evaluate several types of PM welfare effects, as follows.
2 • Vegetation and ecosystem effects.
3 • Visibility effects.
4 • Materials damage.
5 • Role of PM in atmospheric radiative transfer and potential consequences for penetration of
6 biologically harmful UVB to the earth's surface and for climate change.
7 (6) Background Information Topics Useful in Evaluating Health Risks.
8 • New monitoring methods, especially methods used in epidemiology studies.
9 • Indicator topics such as PM2 5 versus PMLO; ultrafine; and PM2 5 versus PM10_2 5.
10 • New data patterns of daily and annual concentrations for PM25, PM10_25, and PM10.
11
12
13 1.4 DOCUMENT ORGANIZATION AND CONTENT
14 The present draft document attempts to critically review and assess relevant scientific
15 literature on PM mainly through December 2001, but does include some more recent studies
16 published in early 2002. The material selected for review and comment in the text generally
17 comes from the more recent literature published since early 1996, with emphasis on studies
18 conducted at or near PM pollutant concentrations found in ambient air. Literature discussed in
19 detail in the previous 1996 EPA PM AQCD (U.S. Environmental Protection Agency, 1996a)
20 generally is not discussed in depth in this document. However, some limited treatment is
21 included of the earlier studies judged to be potentially useful in deriving PM NAAQS. Key
22 literature is presented mainly in tables; and the text mainly discusses overall interpretive points.
23 Primary emphasis is placed on consideration of published material that has undergone
24 scientific peer review. However, in the interest of admitting new and important information
25 expected to become available shortly, some material not yet fully published in the open literature
26 but meeting other standards of scientific reporting (i.e., peer review, quality assurance) are now
27 provisionally included. As noted earlier, emphasis has been placed on studies in the range of
28 current ambient levels. However, studies examining effects of higher concentrations have been
29 included if they contain unique data or documentation of a previously unreported effect or
30 mechanism. In reviewing and summarizing the literature, an attempt has been made to present
31 alternative points of view where scientific controversy exists.
April 2002 1-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 The present document is basically organized to assess information related to topics along
2 the same flow of issues presented in the NRC conceptual framework shown in Figure 1-1.
3 It includes the Executive Summary and nine chapters presented in two volumes. Volume 1
4 contains the Executive Summary and this general introduction (Chapter 1). It also includes
5 Chapters 2 and 3, which provide background information on physical and chemical properties of
6 PM and related compounds; sources and emissions; atmospheric transport, transformation, and
7 fate of PM; methods for the collection and measurement of PM; and U.S. ambient air PM
8 concentrations. Next, Chapter 4 assesses PM environmental effects on vegetation and
9 ecosystems, visibility, man-made materials, and climate-related effects (including effects on solar
10 radiation), and includes limited information on economic impacts of some such welfare effects.
11 Also included in Volume 1 is Chapter 5, which discusses factors affecting exposure of the
12 general population to ambient PM.
13 The second volume contains Chapters 6 through 9. Chapters 6 evaluates information
14 concerning dosimetry of inhaled particles in the respiratory tract. Chapter 7 assesses the
15 toxicology of specific types of PM constituents and potential mechanisms of action, based on
16 both laboratory animal studies and controlled human exposure studies. Chapter 8 discusses
17 observational, i.e., epidemiological, studies. Lastly, Chapter 9 integrates key information on
18 exposure, dosimetry, and critical health risk issues derived from studies reviewed in the prior
19 chapters. That Integrative Synthesis chapter is basically organized in a manner to address the
20 series of 10 issues (and, where appropriate, subissues) identified in the NRC PM Research
21 Priorities Reports (National Research Council, 1998, 1999, 2001).
22
23
April 2002 1-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30)
3 36:8186-8201.
4 Federal Register. (1987) Revisions to the national ambient air quality standards for paniculate matter. F. R. (July 1)
5 52:24,634-24,669.
6 Federal Register. (1996) National ambient air quality standards for paniculate matter; proposed rule. F. R.
7 (November 27) 61: 65638.
8 Federal Register. (1997a) National ambient air quality standards for paniculate matter; final rule. F. R. (July 18)
9 62: 38,652-38,752.
10 Federal Register. (1997b) National ambient air quality standards for ozone; final rule. F. R. (July 18)
11 62:38,856-38,896.
12 Federal Register. (1997c) Implementation of revised air quality standards for ozone and paniculate matter. F. R.
13 (July 18) 62: 38,421-38,422.
14 National Research Council. (1998) Research priorities for airborne paniculate matter. I. Immediate pnorities and a
15 long-range research portfolio. Washington, DC: National Academy Press.
16 National Research Council. (1999) Research priorities for airborne paniculate matter. II. Evaluating research
17 progress and updating the portfolio. Washington, DC: National Academy Press.
18 National Research Council. (2001) Research priorities for airborne paniculate matter. III. Early research progress.
19 Washington, DC: National Academy Press.
20 U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air
21 quality standards. U. S. C. 42: §§7408-7409.
22 U.S. District Court of Arizona. (1995) American Lung Association v. Browner. West's Federal Supplement 884
23 F.Supp. 345 (No. CIV 93-643 TUC ACM).
24 U.S. Environmental Protection Agency. (1996a) Air quality criteria for paniculate matter. Research Triangle Park,
25 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
26 U.S. Environmental Protection Agency. (1996b) Review of the national ambient air quality standards for paniculate
27 matter: policy assessment of scientific and technical information. OAQPS staff paper. Research Triangle
28 Park, NC: Office of Air Quality Planning and Standards; report no. EPA/452/R-96-013. Available from:
29 NTIS, Springfield, VA; PB97-115406REB.
30 U.S. Environmental Protection Agency. (1998a) Paniculate matter research needs for human health risk assessment
31 to support future reviews of the national ambient air quality standards for paniculate matter. Research
32 Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-97/132F.
33 U.S. Environmental Protection Agency. (1998b) Review of national ambient air quality standards for paniculate
34 matter. Commer. Bus. Daily: February 19. Available: http://cbdnet.access.gpo.gov/index.html [1999,
35 November 24].
3 6 Wolff, G. T. (1996) Closure by the Clean Air Scientific Advisory Committee (CASAC) on the staff paper for
37 paniculate matter [letter to Carol M. Browner, Administrator, U.S. EPA]. Washington, DC: U.S.
3 8 Environmental Protection Agency, Clean Air Scientific Advisory Committee;
39 EPA-SAB-CASAC-LTR-96-008; June 13.
40
April 2002 1-19 DRAFT-DO NOT QUOTE OR CITE
-------
i 2. PHYSICS, CHEMISTRY, AND MEASUREMENT
2 OF PARTICULATE MATTER
3
4
5 An extensive review of the physics and chemistry of particulate matter (PM) was included
6 in Chapter 3 of the 1996 EPA document Air Quality Criteria for Parti culate Matter (U.S.
7 Environmental Protection Agency, 1996a). Chapter 2 of this new version of the PM Air Quality
8 Criteria Document (PM AQCD) provides background information on the physics and chemistry
9 of atmospheric particles that may be useful in reading subsequent sections and chapters. This
10 PM AQCD follows the Risk Assessment Framework, as described in Chapter 1 (Section 1.2.2),
11 beginning with sources and continuing to effects as shown in Figure 1-1. However, particular
12 matter, unlike the other criteria pollutants (O3, CO, NO2, and Pb), is not a specific chemical entity
13 but is a mixture of particles of different sizes, compositions, and properties. Therefore, it will be
14 useful to present some background on the chemistry and physics of PM before entering the Risk
15 Assessment Framework. This chapter includes new information that should be useful in
16 understanding risk assessments of the effects of PM on human health and welfare. Emphasis is
17 placed on differences between fine and coarse particles and differences between the nuclei mode
18 and the accumulation mode within fine particles. PM information important for implementation
19 of a standard, but not essential to the standard setting process, is not covered in this chapter. The
20 reader is referred to the NARSTO Fine Particle Assessment (NARSTO, 2002) for information
21 relevant to air quality management for PM.
22 PM is defined quantitatively by the measurement techniques. Therefore, before entering
23 the Risk Assessment Framework, it will also be useful to discuss our understanding of the
24 relationship between PM suspended in the atmosphere, PM inhaled by people, and PM measured
25 by various sampling and analytical techniques. Chapter 4 of the 1996 PM AQCD (U.S.
26 Environmental Protection Agency, 1996a) contained a review of the state-of-the-art of PM
27 measurement technology. Since that time, considerable progress has been made in understanding
28 problems in the measurement of PM mass, chemical composition, and physical parameters.
29 There also has been some progress in developing new and improved measurement techniques,
30 especially for continuous measurements. Therefore, a more extensive survey on measurement
April 2002 2-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 problems and on newly developed measurement techniques is included in Section 2.2. For more
2 detail and older references, the reader is referred to Chapters 3 and 4 of the 1996 PM AQCD
3 (U.S. Environmental Protection Agency, 1996a).
4
5
6 2.1 PHYSICS AND CHEMISTRY OF PARTICULATE MATTER
7 2.1.1 Basic Concepts
8 Atmospheric particles originate from a variety of sources and possess a range of
9 morphological, chemical, physical, and thermodynamic properties. Examples include
10 combustion-generated particles, such as diesel soot or fly ash; photochemically produced
11 particles, such as those found in urban haze; salt particles formed from sea spray; and soil-like
12 particles from resuspended dust. Some particles are liquid; some are solid. Others may contain a
13 solid core surrounded by liquid. Atmospheric particles contain inorganic ions, metallic
14 compounds, elemental carbon, organic compounds, and crustal compounds. Some atmospheric
15 particles are hygroscopic and contain particle-bound water. The organic fraction is especially
16 complex, containing hundreds of organic compounds. (Note: See Appendix 3C for information
17 on the composition of the organic fraction and the concentration of specific organic compounds.)
18 Primary particles are emitted directly from sources. Secondary particles are formed from gases
19 through chemical reactions in the atmosphere involving atmospheric oxygen (O2) and water
20 vapor (H2O); reactive species such as ozone (O3); radicals such as the hydroxyl (»OH) and nitrate
21 (*NO3) radicals; and pollutants such as sulfur dioxide (SO2), nitrogen oxides (NOX), and organic
22 gases from natural and anthropogenic sources. The particle formation process includes
23 nucleation of particles from low-vapor pressure gases emitted from sources or formed in the
24 atmosphere by chemical reactions, condensation of low-vapor pressure gases on existing
25 particles, and coagulation of particles. Thus, any given particle may contain PM from many
26 sources.
27 The composition and behavior of particles are fundamentally linked with those of the
28 surrounding gas. Aerosol may be defined as a suspension of solid or liquid particles in air. The
29 term aerosol includes both the particles and all vapor or gas phase components of air. However,
April 2002 2-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 the term aerosol is sometimes used to refer to the suspended particles only. In this document,
2 "particulate" is used only as an adjective, as in particulate matter.
3 A complete description of the atmospheric aerosol would include an accounting of the
4 chemical composition, morphology, and size of each particle and the relative abundance of each
5 particle type as a function of particle size (Friedlander, 1970). However, the physical and
6 chemical characteristics of particles are usually measured separately. Size distributions by
7 particle number, used to calculate surface area and volume distributions, often are determined by
8 physical means, such as electrical mobility or light scattering of suspended particles. Chemical
9 composition usually is determined by analysis of collected samples although some species can be
10 measured in situ. The mass and average chemical composition of particles, segregated according
11 to aerodynamic diameter by cyclones or impactors, can also be determined. However, recent
12 developments in single particle analysis techniques, by electron microscopy with X-ray analysis
13 of single particles (but not agglomerates) collected on a substrate or by mass spectroscopy of
14 suspended particles passing through a sensing volume, provide elemental composition of
15 individual particles by particle size and, thus, are bringing the description envisioned by
16 Friedlander (1970) closer to reality.
17
18 2.1.2 Physical Properties and Processes
19 2.1.2.1 Definitions of Particle Diameter
20 The diameter of a spherical particle may be determined by optical or electron microscopy,
21 by light scattering and Mie theory, by its electrical mobility, or by its aerodynamic behavior.
22 However, atmospheric particles often are not spherical and vary in density. Therefore, their
23 diameters are often described by an "equivalent" diameter (i.e., that of a unit density sphere that
24 would have the same physical behavior). The aerodynamic diameter is important for particle
25 transport, collection, and respiratory tract deposition. The aerodynamic diameter (Da) depends on
26 the density of the particle. It is defined as the diameter of a spherical particle with a density of
27 1 g/cm3 but with a settling velocity equal to that of the particle in question. Consequently,
28 particles with the same physical size and shape but different densities will have different
29 aerodynamic diameters. Detailed definitions of the various sizes and their relationships are given
30 in standard aerosol textbooks (e.g., Friedlander [1977], Reist [1984, 1993], Seinfeld and Pandis
April 2002 2-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 [1998], Hinds [1999], Vincent [1989, 1995], Willeke and Baron [1993], Baron and Willeke
2 [2001], and Fuchs [1964, 1989]).
3
4 2.1.2.2 Aerosol Size Distributions
5 Particle size, as indexed by one of the "equivalent" diameters, is an important parameter in
6 determining the properties, effects, and fate of atmospheric particles. The atmospheric
7 deposition rates of particles, and therefore, their residence times in the atmosphere, are a strong
8 function of their aerodynamic diameters. The aerodynamic diameter also influences deposition
9 patterns of particles within the lung. Because light scattering is strongly dependent on the optical
10 particle size, the amount of light scattering per unit PM mass will be dependent on the size
11 distribution of atmospheric particles. Therefore, the effects of atmospheric particles on visibility,
12 radiative balance, and climate will be influenced by the size distribution of the particles. Studies
13 using impactors or cyclones measure the particle-size distribution directly in aerodynamic
14 diameter. The diameters of atmospheric particles range from 1 nm to 100 //m, spanning 5 orders
15 of magnitude. A variety of different instruments, measuring a variety of equivalent diameters,
16 are required to cover this range.
17 Older particle counting studies used optical particle counters to cover the range of 0.3 to
18 30 (j,m diameter. Diameters of particles below 0.5 //m were measured as mobility diameters.
19 The particle diameters used in size distribution graphs from these studies usually are given as
20 physical diameters rather than aerodynamic diameters. In recent years, aerodynamic particle
21 sizers have been developed that give a direct measurement of the aerodynamic diameter in the
22 range of approximately 0.7 to 10 //m diameter. These instruments have been used with electrical
23 mobility analyzers that measure the mobility diameter of particles from 3-5 nm to approximately
24 0.5 //m (McMurry, 2000). Unfortunately, there is no agreed-upon technique for combining the
25 various equivalent diameters. Some workers use various assumptions to combine the various
26 measurements into one presentation; others report each instrument separately. Therefore, the
27 user of size distribution data should be careful to determine exactly which equivalent diameter is
28 reported. Aerodynamic diameter is the most widely used equivalent diameter. In this document
29 Dp will be used for physical diameter and Da for aerodynamic diameter.
30
April 2002 2-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 Particle Size Distribution Functions
2 The distribution of particles with respect to size is an important physical parameter
3 governing their behavior. Because atmospheric particles cover several orders of magnitude in
4 particle size, size distributions often are expressed in terms of the logarithm of the particle
5 diameter on the X-axis and the measured differential concentration on the Y-axis: AN/A(logDp)
6 = the number of particles per cm3 of air having diameters in the size range from log Dp to log(Dp
7 + ADp). Because logarithms do not have dimensions, it is necessary to think of the distribution
8 as a function of log(Dp/Dp0), where the reference diameter Dp0 = 1 //m is not explicitly stated. If
9 AN/A(logDp) is plotted on a linear scale, the number of particles between Dp and Dp + ADp is
10 proportional to the area under the curve of AN/A(logDp) versus logDp. Similar considerations
11 apply to distributions of surface, volume, and mass. It has been found that atmospheric aerosol
12 size distributions frequently may be approximated by a sum of log-normal distributions
13 corresponding to the various modes or fractions. When approximated by a function, the
14 distributions are usually given as dN/d(log Dp) rather than AN/A(log Dp).
15
16 Atmospheric Aerosol Size Distributions
17 Averaged atmospheric size distributions are shown in Figures 2-1 through 2-3 (Whitby,
18 1978; Whitby and Sverdrup, 1980). Figure 2-1 describes the number of particles as a function of
19 particle diameter for rural, urban-influenced rural, urban, and freeway-influenced urban aerosols.
20 For some of the same data, the particle volume distributions are shown in Figure 2-2. Figure 2-3
21 shows the number, surface, and volume distribution for the grand average continental size
22 distribution. Volume, surface area, and sometimes number are shown on an arithmetic scale with
23 the distributions plotted such that the volume, surface area, or number of particles in any
24 specified size range is proportional to the corresponding area under the curve. These
25 distributions show that most of the particles are quite small, below 0.1 //m; whereas most of the
26 particle volume (and therefore most of the mass) is found in particles >0.1 //m.
27 An important feature of the mass or volume size distributions of atmospheric aerosols is
28 their multimodal nature. Volume distributions, measured in ambient air in the United States, are
29 almost always found to be bimodal with a minimum between 1 and 3 //m. The distribution of
30 particles that are mostly larger than the minimum is termed "coarse." The distribution of
31 particles that are mostly smaller than the minimum is termed "fine." Whitby and Sverdrup
April 2002 2-5 DRAFT-DO NOT QUOTE OR CITE
-------
1,000,000 -
10,000 -
1-
0.01 H
0.0001 -
0.000001 -
\ r
Clean Rural
Urban Influenced Rural
Average Urban
Urban + Freeway
200,000
0.01 0.1 1 10 100
Particle Diameter, Dp (|jm)
150,000
O
_o
T3
100,000 -
50,000 -
0.01 0.1 1 10
Particle Diameter, Dp (|jm)
Figure 2-1. Number of particles as a function of particle diameter: (a) number
concentrations are shown on a logarithmic scale to display the wide range by
site and size; (b) number concentrations for the average urban distribution are
shown on a linear scale for which the area under any part of the curve is
proportional to particle number in that size range.
Source: Whitby and Sverdrup (1980).
1 (1980), Whitby (1978), and Willeke and Whitby (1975) identified three modes: (1) nuclei,
2 (2) accumulation, and (3) coarse. The three modes are most apparent in the freeway-influenced
3 size distribution of Figure 2-2b, in the surface area distribution of Figure 2-3b, and in the
4 in-traffic volume distribution of Figure 2-4. However, the nuclei mode, corresponding to
5 particles below about 0.1 //m, may not be noticeable in volume or mass distributions. The
6 middle mode, from 0.1 to 1 or 2 //m, is the accumulation mode. Fine particles include both the
7 accumulation and the nuclei modes. The third mode, containing particles larger than 1 or 2 //m,
8 is known as the coarse particle mode. The number concentrations of coarse particles are usually
9 too small to be seen in arithmetic plots (Figures 2-lb and 2-3a) but can be seen in a logarithmic
April 2002
2-6
DRAFT-DO NOT QUOTE OR CITE
-------
n.
Q
D)
%
1
/u -
65 -
60 -
55 -
50 -
35 -
30 -
25 -
20 -
15 -
10 -
5 -
o H
Q ,•, Clean Rural
• \
• \ Rural
.' ; South-Central
: ': New Mexico
/ v A\
/ / \\
/ ^x""x^X
. .j s ^ |»-/ ! ,^..X
70
65
60 -
55 -
50 -
^40 H
^35 H
"3
9. 30 -
O)
=§ 25 '
^ 20 -
15 -
10 -
5 -
0
Average Urban
Urban + Freeway
0.01 0.1 1 10
Particle Diameter, Dp (pm)
100 0.01 0.1 1 10
Particle Diameter, Dp (|jm)
100
Figure 2-2. Particle volume distribution as a function of particle diameter: (a) for the
averaged rural and urban-influenced rural number distributions shown in
Figure 2-1 and a distribution from south central New Mexico, and (b) for the
averaged urban and freeway-influenced urban number distributions shown in
Figure 2-1.
Source: Whitby and Sverdrup (1980) and Kim et al. (1993).
1 plot (Figure 2-la). Whitby and Sverdrup (1980) observed that rural aerosols, not influenced by
2 nearby sources, have a small accumulation mode and no observable nuclei mode. For urban
3 aerosols, the accumulation and coarse particle modes are comparable in volume. The nuclei
4 mode is small in volume, but it dominates the number distributions of urban aerosols. Whitby's
5 conclusions were based on extensive studies of size distributions in a number of western and
6 midwestern locations during the 1970s (Whitby, 1978; Whitby and Sverdrup, 1980).
7 No size-distribution studies of similar scope have been published since then. Newer results from
8 particle counting and impactor techniques, including data from Europe (U.S. Environmental
9 Protection Agency, 1996a) and Australia (Keywood et al., 1999, 2000), show similar results.
April 2002
2-7
DRAFT-DO NOT QUOTE OR CITE
-------
CD
_Q
E
CO
^
CO
X
CO
o
Q
O)
o
a.
Q
O)
o
c/
15-
10 -
5-
600-
400-
200-
30 H
(l)
E
Q
O)
20 H
10-
Nn = 7.7x 10
DGNn = 0.013
agn=1.7
Na = 1.3x 10
Nc=4.2
DGNC= 0.97
agc=2.15
i | 11111
Sa = 535
DGSa = 0.19
DGSn = 0.023
Sc=41
DGSC= 3.1
Vn = 0.33
DGVn = 0.031
0.001
• I I I I l|
0.01
(a)
(b)
0.1
1.0
10
100
Figure 2-3. Distribution of coarse (c), accumulation (a), and nuclei- or ultrafine (n) -mode
particles by three characteristics, a) number (N), b) surface area (S) and
c) volume (V) for the grand average continental size distribution. DGV =
geometric mean diameter by volume; DGS = geometric mean diameter by
surface area; DGN = geometric mean diameter by number; Dp = particle
diameter.
Source: Whitby (1978).
April 2002
2-8
DRAFT-DO NOT QUOTE OR CITE
-------
o
5 -
D
O)
3 -
2 -
1 -
Vapor
Mechanically
Generated
Nucleation
DGV = 0.018
a =1.6
0.002
0.01
Nuclei Mode
0.1 1
Particle Diameter, Dp(|jm)
Accumulation Mode
10
Fine-Mode Particles
Coarse Mode
Coarse-Mode Particles
100
Figure 2-4. Volume size distribution, measured in traffic, showing fine-mode and
coarse-mode particles and the nuclei and accumulation modes within the
fine-particle mode. DGV (geometric mean diameter by volume, equivalent to
volume median diameter) and og (geometric standard deviation) are shown for
each mode. Also shown are transformation and growth mechanisms (e.g.,
nucleation, condensation, and coagulation).
Source: Adapted from Wilson and Suh (1997).
1 Definitions of Particle Size Fractions
2 In the preceding discussion several subdivisions of the aerosol size distribution were
3 identified. Aerosol scientists use four different approaches or conventions in the classification of
4 particles by size: (1) modes, based on the observed size distributions and formation mechanisms;
5 (2) cut point, usually based on the 50% cut point of the specific sampling device; (3) dosimetry
6 or occupational health sizes, based on the entrance into various compartments of the respiratory
7 system; and (4) legally specified, regulatory sizes for air quality standards.
April 2002
2-9
DRAFT-DO NOT QUOTE OR CITE
-------
1 Modal. The modal classification, first proposed by Whitby (1978), is shown in Figure 2-3.
2 The nuclei mode can be seen clearly in the volume distribution only in traffic or near traffic or
3 other sources of nuclei mode particles (Figure 2-4). The observed modal structure is frequently
4 approximated by several log-normal distributions. Definitions of terms used to describe size
5 distributions in modal terms are given below.
6
7 Coarse Mode: The distribution of particles with diameters mostly greater than the
8 minimum in the particle mass or volume distributions, which generally occurs between
9 1 and 3 //m. These particles are usually mechanically generated (e.g., from wind erosion of
10 crustal material).
11
12 Fine Mode: The distribution of particles with diameters mostly smaller than the minimum
13 in the particle mass or volume distributions, which generally occurs between 1 and 3 //m.
14 These particles are generated in combustion or formed from gases. The fine mode includes
15 the accumulation mode and the nuclei mode.
16
17 Nuclei Mode: That portion of the fine particle mode with diameters below about 0.1 //m.
18 The nuclei mode can be observed as a separate mode in mass or volume distributions only
19 in clean or remote areas or near sources of new particle formation by nucleation.
20 Toxicologists and epidemiologists use the term "ultrafme" to refer to particles in the
21 nuclei-mode size range. Aerosol physicists and material scientists tend to use the term
22 "nanoparticles" to refer to particles in this size range generated in the laboratory.
23
24 Accumulation Mode: That portion of the fine particle mode with diameters above about
25 0.1 //m. Accumulation-mode particles normally do not grow into the coarse mode.
26 Nuclei-mode particles grow by coagulation (two particles combining to form one) or by
27 condensation (low-equilibrium vapor pressure gas molecules condensing on a particle) and
28 "accumulate" in this size range.
29
30 Over the years, the terms fine and coarse, as applied to particle sizes, have lost the precise
31 meaning given in Whitby's (1978) definition. In any given article, therefore, the meaning of fine
April 2002 2-10 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
and coarse, unless defined, must be inferred from the author's usage. In particular, PM2 5 and
fine-mode particles are not equivalent. In this document, the term "mode" is used with fine and
coarse when it is desired to specify the distribution of fine-mode particles or coarse-mode
particles as shown in Figures 2-4 and 2-5.
70
:o
CD
CL
en
.o
V)
60 -
u 50
0
CD
Q 40
Q)
o
30 -
20 -
10 -
Fine-Mode Particles
Coarse-Mode Particles
0.1 0.2
0.5 1.0 2 5 10 20
Aerodynamic Particle Diameter (|jm)
Total Suspended Particles (TSP)
PW
10
PM
2.5
PM
10-2.5
TSP
HiVol
WRAC
50 100
Figure 2-5. An idealized distribution of ambient particulate matter showing fine-mode
particles and coarse-mode particles and the fractions collected by size-selective
samplers. (WRAC is the Wide Range Aerosol Classifier which collects the
entire coarse mode [Lundgren and Burton, 1995].)
Source: Adapted from Wilson and Suh (1997).
April 2002
2-11
DRAFT-DO NOT QUOTE OR CITE
-------
1 Sampler Cut Point. Another set of definitions of particle size fractions arises from
2 considerations of size-selective sampling. Size-selective sampling refers to the collection of
3 particles below or within a specified aerodynamic size range, usually defined by the upper 50%
4 cut point size, and has arisen in an effort to measure particle size fractions with some special
5 significance (e.g., health, visibility, source apportionment, etc.). Dichotomous samplers split the
6 particles into smaller and larger fractions that may be collected on separate filters. However,
7 some fine particles (~ 10%) are collected with the coarse particle fraction. Cascade impactors use
8 multiple size cuts to obtain a distribution of size cuts for mass or chemical composition
9 measurements. One-filter samplers with a variety of upper size cuts also have been used.
10
11 Occupational Health or Dosimetric Size Cuts. The occupational health community has
12 defined size fractions for use in the protection of human health. This convention classifies
13 particles into inhalable, thoracic, and respirable particles according to their upper size cuts.
14 However, these size fractions may also be characterized in terms of their entrance into various
15 compartments of the respiratory system. Thus, inhalable particles enter the respiratory tract,
16 including the head airways. Thoracic particles travel past the larynx and reach the lung airways
17 and the gas-exchange regions of the lung. Respirable particles are a subset of thoracic particles
18 that are more likely to reach the gas-exchange region of the lung. In the past exact definitions of
19 these terms have varied among organizations. As of 1993, a unified set of definitions was
20 adopted by the American Conference of Governmental Industrial Hygienists (ACGIH, 1994), the
21 International Standards Organization (ISO), and the European Standardization Committee
22 (CEN). The curves which define inhalable (IPM), thoracic (TPM), and respirable (RPM)
23 particulate matter are shown in Figure 2-6.
24
25 Regulatory Size Cuts In 1987, the NAAQS for PM were revised to use PM10, rather than
26 total suspended particulate matter (TSP), as the indicator for the NAAQS for PM (Federal
27 Register, 1987). The use of PM10 as an indicator is an example of size-selective sampling based
28 on a regulatory size cut (Federal Register, 1987). The selection of PM10 as an indicator was
29 based on health considerations and was intended to focus regulatory concern on those particles
30 small enough to enter the thoracic region of the human respiratory tract. The PM2 5 standard set
31 in 1997 is also an example of size-selective sampling based on a regulatory size cut (Federal
April 2002 2-12 DRAFT-DO NOT QUOTE OR CITE
-------
100
0
APM10
• IPM
• TPM
O RPM
VPM25
4 10 20 50
Aerodynamic Diameter (|jm)
100
Figure 2-6. Specified particle penetration (size-cut curves) through an ideal (no-particle-
loss) inlet for five different size-selective sampling criteria. Regulatory size
cuts are defined in the Code of Federal Regulations; PM25 (2001c), PM10
(2001a). PM2 5 is also defined in the Federal Register (1997). Size-cut curves
for inhalable particulate matter (IPM), thoracic particulate matter (TPM) and
respirable particulate matter (RPM) size cuts are computed from definitions
given by American Conference of Governmental and Industrial Hygienists
(1994).
1 Register, 1997). The PM25 standard was based primarily on epidemiological studies using
2 concentrations measured with PM2 5 samplers as an exposure index. However, the PM2 5 sampler
3 was not designed to collect respirable particles. It was designed to collect fine-mode particles
4 because of their different sources (Whitby et al., 1974). Thus, the need to attain a PM2 5 standard
5 will tend to focus regulatory concern on control of sources of fine-mode particles.
6 Prior to 1987, the indicator for the NAAQS for PM was TSP. TSP is defined by the design
7 of the High Volume Sampler (hivol) that collects all of the fine particles but only part of the
April 2002
2-13
DRAFT-DO NOT QUOTE OR CITE
-------
1 coarse particles (Figure 2-5). The upper cut-off size of the hivol depends on the wind speed and
2 direction and may vary from 25 to 40 //m. The Wide Range Aerosol Classifier (WRAC) was
3 designed specifically to collect the entire coarse mode (Lundgren and Burton, 1995).
4 An idealized distribution, showing the normally observed division of ambient aerosols into
5 fine-mode particles and coarse-mode particles and the size fractions collected by the WRAC,
6 TSP, PM10, PM25 and PM10_2 5 samplers, is shown in Figure 2-5. PM10 samplers, as defined in
7 Appendix J to Title 40 Code of Federal Regulations (40 CFR) Part 50 (Code of Federal
8 Regulations, 200 la; Federal Register, 1987), collect all of the fine particles and part of the coarse
9 particles. The upper cut point is defined as having a 50% collection efficiency at 10 ± 0.5 //m
10 aerodynamic diameter. The slope of the collection efficiency curve is defined in amendments to
11 40 CFR, Part 53, (Code of Federal Regulations, 2001b). An example of a PM10 size-cut curve is
12 shown in Figure 2-6.
13 An example of a PM25 size-cut curve is also shown in Figure 2-6. The PM25 size-cut
14 curve, however, is defined by the design of the Federal Reference Method (FRM) Sampler. The
15 basic design of the FRM is given in the Federal Register (1997, 1998) and as 40 CFR Part 50,
16 Appendix L (Code of Federal Regulations, 200 Ic). Additional performance specifications are
17 given in 40 CFR Parts 53 and 58 (Code of Federal Regulations, 2001b,d). Each actual PM2 5
18 reference method, as represented by a specific sampler design and associated manual operational
19 procedures, must be designated as a reference method under 40 CFR Part 53 in Section 1.2 of
20 Appendix L (Code of Federal Regulations, 200Ic). Thus there may be many somewhat different
21 PM2 5 FRMs (see Table 2-4).
22 Papers discussing PM10 or PM2 5 frequently insert an explanation such as "PMX (particles
23 less than x //m diameter)" or "PMX (nominally, particles with aerodynamic diameter x are collected and not all particles < x are collected.
April 2002 2-14 DRAFT-DO NOT QUOTE OR CITE
-------
O
U— <
-------
1 The resulting PM10_2 5 mass, or PM10_2 5, is sometimes called "coarse" particles. However,
2 it would be more correct to call PM25 an indicator of fine-mode particles (because it contains
3 some coarse-mode particles) and PM10_2 5 an indicator of the thoracic component of coarse-mode
4 particles (because it excludes some coarse-mode particles below 2.5 //m Da and above 10 //m
5 Da). It would be appropriate to call PM10 an indicator of thoracic particles. PM10 and thoracic
6 PM, as shown in Figure 2-6, have the same 50% cut point. However, the thoracic cut is not as
7 sharp as the PM10 cut; therefore, thoracic PM contains some particles between 10 and 30 //m
8 diameter that are excluded from PM10.
9
10 2.1.2.3 Nuclei-Mode Particles
11 As discussed in Chapter 7, Toxicology of Particulate Matter, and in Chapter 8,
12 Epidemiology of Human Health Effects from Ambient Parti culate Matter, some scientists argue
13 that ultrafine (nuclei-mode) particles may pose potential health problems and that some health
14 effects may be more closely associated with particle number or particle surface area than particle
15 mass. Because nuclei-mode particles contribute the major portion of particle number and a
16 significant portion of particle surface area, some additional attention will be given to
17 nuclei-mode particles.
18
19 Formation and Growth of Fine Particles
20 Several processes influence the formation and growth of particles. New particles may be
21 formed by nucleation from gas phase material. Particles may grow by condensation as gas phase
22 material condenses on existing particles. Particles also may grow by coagulation as two particles
23 combine to form one. Gas phase material condenses preferentially on smaller particles, and the
24 rate constant for coagulation of two particles decreases as the particle size increases. Therefore,
25 nuclei mode particles grow into the accumulation mode, but accumulation mode particles do not
26 normally grow into the coarse mode (see Figure 2-4). More information and references on
27 formation and growth of fine particles may be found in the 1996 AQCD PM (U.S. Environmental
28 Protection Agency, 1996a).
29
30
April 2002 2-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 Equilibrium Vapor Pressures
2 An important parameter in particle nucleation and in particle growth by condensation is the
3 saturation ratio S, defined as the ratio of the partial pressure of a species, p, to its equilibrium
4 vapor pressure above a flat surface, p0: S = p/p0. For either condensation or nucleation to occur,
5 the species vapor pressure must exceed its equilibrium vapor pressure. For particles, the
6 equilibrium vapor pressure is not the same as p0. Two effects are important: (1) the Kelvin
7 effect, which is an increase in the equilibrium vapor pressure above the surface due to its
8 curvature (very small particles have higher vapor pressures and will not be stable to evaporation
9 until they attain a critical size) and (2) the solute effect, which is a decrease in the equilibrium
10 vapor pressure of the liquid due to the presence of other compounds in solution. Organic
11 compounds may also be adsorbed on ultrafme carbonaceous particles.
12 For an aqueous solution of a nonvolatile salt, the presence of the salt decreases the
13 equilibrium vapor pressure of the water over the droplet. This effect is in the opposite direction
14 of the Kelvin effect, which increases the equilibrium vapor pressure above a droplet because of
15 its curvature. The existence of an aqueous solution will also influence the vapor pressure of
16 water-soluble species. The vapor pressure behavior of mixtures of several liquids or of liquids
17 containing several solutes is complex.
18
19 New Particle Formation
20 When the vapor concentration of a species exceeds its equilibrium concentration (expressed
21 as its equilibrium vapor pressure), it is considered condensable. Condensable species can either
22 condense on the surface of existing particles or can form new particles. The relative importance
23 of nucleation versus condensation depends on the rate of formation of the condensable species
24 and on the surface or cross-sectional area of existing particles (McMurry and Friedlander, 1979).
25 In ambient urban environments, the available particle surface area is sufficient to rapidly
26 scavenge the newly formed condensable species. Formation of new particles (nuclei mode) is
27 usually not important except near sources of condensable species. Wilson et al. (1977) report
28 observations of the nuclei mode in traffic. New particle formation also can be observed in
29 cleaner, remote regions. Bursts of new particle formation in the atmosphere under clean
30 conditions usually occur when aerosol surface area concentrations are low (Covert et al., 1992).
31 High concentrations of nuclei mode particles have been observed in regions with low particle
April 2002 2-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 mass concentrations indicating that new particle formation is inversely related to the available
2 aerosol surface area (Clarke, 1992).
3
4 Sources of Nuclei-Mode Particles
5 Nuclei mode particles are the result of nucleation of gas phase species to form condensed
6 phase species with very low equilibrium vapor pressure. In the atmosphere there are four major
7 classes of sources that yield particulate matter with equilibrium vapor pressures low enough to
8 form nuclei mode particles:
9 (1) Particles containing heavy metals. Nuclei mode particles of metal oxides or other
10 metal compounds are generated when metallic impurities in coal or oil are vaporized during
11 combustion and the vapor undergoes nucleation. Metallic ultrafine particles also may be
12 formed from metals in lubricating oil or fuel additives that are vaporized during
13 combustion of gasoline or diesel fuels. Nuclei-mode metallic particles were discussed in
14 Section 6.9 of the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a).
15 (2) Elemental carbon or soot (EC). EC particles are formed primarily by condensation of
16 C2 molecules generated during the combustion process. Because EC has a very low
17 equilibrium vapor pressure, ultrafine EC particles can nucleate even at high temperatures
18 (Kittelson, 1998; Morawska et al., 1998).
19 (3) Sulfates and nitrates. Sulfuric acid (H2SO4), or its neutralization products with
20 ammonia (NH3), ammonium sulfate ((NH4)2SO4) or ammonium acid sulfate (NH4HSO4),
21 are generated in the atmosphere by conversion of sulfur dioxide (SO2) to H2SO4. As H2SO4
22 is formed, it can either nucleate to form new ultrafine particles, or it can condense on
23 existing nuclei mode or accumulation mode particles (Clark and Whitby, 1975; Whitby,
24 1978). The possible formation of ultrafine NH4NO3 by reaction of NH3 and nitric acid
25 (HNO3) vapor apparently has not been investigated.
26 (4) Organic carbon. Recent smog chamber studies and indoor experiments show that
27 atmospheric oxidation of certain organic compounds found in the atmosphere can produce
28 highly oxidized organic compounds with an equilibrium vapor pressure sufficiently low to
29 result in nucleation (Kamens et al., 1999; Weschler and Shields, 1999).
30
31
April 2002 2-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 Recent Measurements of Nuclei-Mode Particles
2 Instruments, developed during the past decade, permit measurement of size distributions
3 down to 3 nm diameter particles. Use of these techniques have led to new information on the
4 formation of new particles by nucleation. Such measurements have been carried out during
5 intensive field measurement campaigns and also during continuous measurements in urban areas
6 in several European cities and in the U.S. as a part of the Supersite program (McMurry et. al.,
7 2000; Woo et al., 2001a). Nucleation has been observed in the outflows of convective clouds,
8 downwind of coastal regions during low tide, over forests, downwind of certain biogenic
9 emissions, and in urban areas. Nucleation events in outdoor air almost always occur during
10 daylight, indicating that photochemistry plays a role in producing the gas phase precursors of
11 new particles. There is strong evidence that sulfuric acid vapor often participates in nucleation.
12 However, condensation of sulfuric acid and its associated water and ammonia typically can
13 account for only 10% to 20% of the observed growth rates for freshly nucleated particles.
14 Therefore, organic compounds may account for much of the formation and growth of freshly
15 nucleated particles. Evidence of nucleation of organic particles comes largely from smog
16 chamber studies (Kamens et al., 1999). Nucleation of organic particles may also occur indoors
17 due to the reaction of infiltrated ozone with indoor terpenes from air fresheners or cleaning
18 solutions (Weschler and Shields, 1999). The observation of bursts of nuclei-mode particles in
19 Atlanta (Woo et al., 2001a), perhaps due to unusually high rates of production of condensible
20 species, suggests that exposure to high concentrations of ultrafine or nuclei-mode particles may
21 be a more frequent occurrence that previously expected.
22
23 Concentration of Nuclei-Mode Particles: A Balance Between Formation and Removal
24 Nuclei-mode particles may be removed by dry deposition or by growth into the
25 accumulation mode. This growth takes place as other low vapor pressure material condenses on
26 the particles or as nuclei-mode particles coagulate with themselves or with accumulation mode
27 particles. Because the rate of coagulation would vary with the concentration of accumulation-
28 mode particles, it might be expected that the concentration of nuclei-mode particles would
29 increase with a decrease in accumulation-mode mass. On the other hand, the concentration of
30 particles would be expected to decrease with a decrease in the rate of generation of particles by
31 reduction in emissions of metal and carbon particles or a decrease in the rate of generation of
April 2002 2-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 H2SO4 or condensable organic vapor. The rate of generation of H2SO4 depends on the
2 concentration of SO2 and hydroxyl radical (•OH), which is generated primarily by reactions
3 involving ozone (O3). Thus, reductions in SO2 and O3 would lead to a decrease in the rate of
4 generation of H2SO4 and condensable organic vapor and to a decrease in the concentration of
5 nuclei-mode particles. The balance between formation and removal is uncertain. However,
6 these processes can be modeled using a general dynamic equation for particle size distribution
7 (Friedlander, 1977) or by aerosol dynamics modules in newer air quality models (Binkowski and
8 Shanker, 1995; Binkowski and Ching, 1995).
9
10 2.1.3 Chemistry of Atmospheric Particulate Matter
11 The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen
12 ions; particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
13 material. Atmospheric PM also contains a large number of elements in various compounds and
14 concentrations. More information and references on the composition of PM, measured in a large
15 number of studies in the United States, may be found in 1996 PM AQCD (U.S. Environmental
16 Protection Agency, 1996a). The composition and concentrations of PM are discussed in
17 Chapter 3, Section 3.1, Patterns and Trends in Ambient PM25 Concentrations. Ambient data for
18 concentrations and composition of PM25 are given in Appendices 3 A, 3B, and 3C.
19
20 2.1.3.1 Chemical Composition and Its Dependence on Particle Size
21 Studies conducted in most parts of the United States indicate that sulfate, ammonium, and
22 hydrogen ions; elemental carbon, secondary organic compounds and primary organic species
23 from cooking and combustion; and certain transition metals are found predominantly in the fine
24 particle mode. Crustal materials such as calcium, aluminum, silicon, magnesium, and iron are
25 found predominately in the coarse particles. Some organic materials such as pollen, spores, and
26 plant and animal debris are also found predominantly in the coarse mode. Some components
27 such as potassium and nitrate may be found in both the fine and coarse particle modes but from
28 different sources or mechanisms. Potassium in coarse particles comes from soil. Potassium also
29 is found in fine particles in emissions from burning wood or cooking meat. Nitrate in fine
30 particles comes primarily from the reaction of gas-phase nitric acid with gas-phase ammonia to
April 2002 2-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 form particulate ammonium nitrate. Nitrate in coarse particles comes primarily from the reaction
2 of gas-phase nitric acid with preexisting coarse particles.
3
4 2.1.3.2 Primary and Secondary Particulate Matter
5 Particulate material can be primary or secondary. PM is called "primary" if it is in the
6 same chemical form in which it was emitted into the atmosphere. PM is called "secondary" if it
7 is formed by chemical reactions in the atmosphere. Primary coarse particles are usually formed
8 by mechanical processes. This includes material emitted in particulate form such as wind-blown
9 dust, sea salt, road dust, and combustion-generated particles such as fly ash and soot. Primary
10 fine particles are emitted from sources either directly as particles or as vapors that rapidly
11 condense to form ultrafme or nuclei-mode particles. This includes soot from diesel engines,
12 a great variety of organic compounds condensed from incomplete combustion or cooking, and
13 compounds of As, Se, Zn, etc., that condense from vapor formed during combustion or smelting.
14 The concentration of primary particles depends on their emission rate, transport and dispersion,
15 and removal rate from the atmosphere.
16 Secondary PM is formed by chemical reactions of free, adsorbed, or dissolved gases. Most
17 secondary fine PM is formed from condensable vapors generated by chemical reactions of
18 gas-phase precursors. Secondary formation processes can result in either the formation of new
19 particles or the addition of particulate material to pre-existing particles. Most of the sulfate and
20 nitrate and a portion of the organic compounds in atmospheric particles are formed by chemical
21 reactions in the atmosphere. Secondary aerosol formation depends on numerous factors
22 including the concentrations of precursors; the concentrations of other gaseous reactive species
23 such as ozone, hydroxyl radical, peroxy radicals, or hydrogen peroxide; atmospheric conditions
24 including solar radiation and relative humidity (RH); and the interactions of precursors and
25 pre-existing particles within cloud or fog droplets or in the liquid film on solid particles. As a
26 result, it is considerably more difficult to relate ambient concentrations of secondary species to
27 sources of precursor emissions than it is to identify the sources of primary particles.
28 A significant effort is currently being directed toward the identification and modeling of organic
29 products of photochemical smog including the conversion of gases to particulate matter. More
30 information of the transformation of precursor gases into secondary PM is given in Chapter 3,
31 Section 3.3.1, Chemistry of Secondary PM Formation.
April 2002 2-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.1.3.3 Particle-Vapor Partitioning
2 Several atmospheric aerosol species, such as ammonium nitrate and certain organic
3 compounds, are semivolatile and are found in both gas and particle phases. A variety of
4 thermodynamic models have been developed to predict the temperature and relative humidity
5 dependence of the ammonium nitrate equilibria with gaseous nitric acid and ammonia. However,
6 under some atmospheric conditions, such as cool, cold, or very clean air, the relative
7 concentrations of the gas and solid phases are not accurately predicted by equilibrium
8 considerations alone, and transport kinetics can be important. The gas-particle distribution of
9 semivolatile organic compounds depends on the equilibrium vapor pressure of the compound,
10 total particle surface area, particle composition, atmospheric temperature, and relative humidity.
11 Although it generally is assumed that the gas-particle partitioning of semivolatile organics is in
12 equilibrium in the atmosphere, neither the equilibria nor the kinetics of redistribution are well
13 understood. Diurnal temperature fluctuations cause gas-particle partitioning to be dynamic on a
14 time scale of a few hours and can cause semivolatile compounds to evaporate during the
15 sampling process. The pressure drop across the filter can also contribute to the loss of
16 semivolatile compounds. The dynamic changes in gas-particle partitioning caused by changes in
17 temperature, pressure, and gas-phase concentration, both in the atmosphere and after collection,
18 cause serious sampling problems that are discussed in Section 2.2.3, Measurement of
19 Semivolatile Particulate Matter.
20
21 Equilibria with Water Vapor
22 As a result of the equilibrium of water vapor with liquid water in hygroscopic particles,
23 many ambient particles contain liquid water (particle-bound water). Unless removed, this
24 particle-bound water will be measured as a component of the particle mass. Particle-bound water
25 is important in that it influences the size of the particles, and in turn, their light scattering
26 properties and their aerodynamic properties, which are important for deposition to surfaces, to
27 airways following inhalation, and in sampling instrumentation. The aqueous solution provides a
28 medium for reactions of dissolved gases including reactions that do not take place in the gas
29 phase. The aqueous solutions also may act as a carrier to convey soluble toxic species to the gas-
30 exchange regions of the respiratory system, including species that would be removed by
31 deposition in the upper airways if the particles had remained in the gas phase (Friedlander and
April 2002 2-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 Yeh, 1998; Kao and Friedlander, 1995; Wilson, 1995). An extensive review of equilibrium with
2 water vapor as it pertains to ambient aerosols was given in Chapter 3 of the 1996 PM AQCD
3 (U.S. Environmental Protection, Agency, 1996a).
4 The interaction of particles with water vapor may be described briefly as follows.
5 As relative humidity increases, particles of crystalline soluble salts, such as (NH4)2SO4,
6 NH4HSO4, or NH4NO3, undergo a phase transition to become aqueous solution particles.
7 According to the phase rule, for particles consisting of a single component, this phase transition
8 is abrupt, taking place at a relative humidity that corresponds to the vapor pressure of water
9 above the saturated solution (the deliquescence point). With a further increase in relative
10 humidity, the solution particle adds water (and the concentration of the solute decreases) so that
11 the vapor pressure of the solution is maintained equal to that of the surrounding relative
12 humidity; thus, the solution particle tends to follow the equilibrium growth curve. As relative
13 humidity decreases, the solution particle follows the equilibrium curve to the deliquescence
14 point. However, rather than crystallizing at the deliquescence relative humidity, the solute
15 remains dissolved in a supersaturated solution to considerably lower relative humidities.
16 Ultimately the solution particle abruptly loses its water vapor (efflorescence) and typically
17 returns to the initial crystalline form.
18 For particles consisting of more than one component, the solid to liquid transition will take
19 place over a range of relative humidities with an abrupt onset at the lowest deliquescence point of
20 the several components and with subsequent growth as crystalline material in the particle
21 dissolves according to the phase diagram for the particular multicomponent system. Under such
22 circumstances, a single particle may undergo several more or less abrupt phase transitions until
23 the soluble material is fully dissolved. At decreasing relative humidity, such particles tend to
24 remain in solution to relative humidities well below the several deliquescence points. In the case
25 of the sulfuric acid-ammonium sulfate-water system, the phase diagram is fairly well worked out.
26 Mixed anion systems containing nitrate are more labile because of the equilibrium between
27 particulate NH4NO3 and gaseous NH3 and HNO3. For particles of composition intermediate
28 between NH4HSO4 and (NH4)2SO4, this transition occurs in the range from 40% to below 10%,
29 indicating that for certain compositions the solution cannot be dried in the atmosphere. At low
30 relative humidities, particles of this composition would likely be present in the atmosphere as
April 2002 2-23 DRAFT-DO NOT QUOTE OR CITE
-------
1 supersaturated solution droplets (liquid particles) rather than as solid particles. Thus, they would
2 exhibit hygroscopic rather than deliquescent behavior during relative humidity cycles.
3 Other pure compounds, such as sulfuric acid, are hygroscopic (i.e., they form aqueous
4 solutions at any relative humidity and maintain a solution vapor pressure over the entire range of
5 relative humidity). Soluble organic compounds may also contribute to the hygroscopicity of the
6 atmospheric aerosol (Saxena et al., 1995; Saxena and Hildeman, 1996), but the equilibria
7 involving organic compounds and water vapor, and, especially for mixtures of salts, organic
8 compounds, and water, are not so well understood. These equilibrium processes may cause an
9 ambient particle to significantly increase its diameter at relative humidities above about 40%
10 (Figure 2-8). A particle can grow to five times its dry diameter as the RH approaches 100%
11 (Figure 2-9). The Federal Reference Methods, for filter measurements of PM2 5 and PM10 mass,
12 require equilibration at a specified, low relative humidity (-40% RH) after collection. This
13 equilibration removes much of the particle-bound water and provides a stable PM mass (see
14 Section 2.2 for details and references). Otherwise, particle mass would be a function of relative
15 humidity, and the particle mass would be largely particle-bound water at higher relative
16 humidities.
17 Continuous monitoring techniques generally attempt to remove particle-bound water before
18 measurement, either by heating or dehumidification. Semivolatile material may be lost during
19 sampling or equilibration; it is certainly lost when the collected sample is heated above ambient
20 temperature. In addition to problems due to the loss of semivolatile species, recent studies have
21 shown that significant amounts of particle-bound water are retained in particles collected on
22 impaction surfaces even after equilibration and that the amount of retained particle-bound water
23 increases with relative humidity during collection (Hitzenberger et al., 1997). Large increases in
24 mass with increasing relative humidity were observed for the accumulation mode. The change in
25 particle size with relative humidity also means that particle measurements such as surface area or
26 volume, or composition as a function of size, should be made at the same RH in order for the
27 results are to be comparable. These problems are addressed below in more detail, in Section 2.2
28 on Measurement of Particulate Matter.
29
April 2002 2-24 DRAFT-DO NOT QUOTE OR CITE
-------
2.0-
CO
01
o
(5
CD
-t— •
CD
CO
b
1.5-
1.0-
Hygroscopic Growth
Curve for (H2SO4)
B
\
30
Hysteresis Loop
for (NH4)2 S04
M
\
50
\
70
RH'°-
8
7
O
-6 >
h 5 O
"co
- 4
- 3
- 2
- 1
90
O
0
E
Figure 2-8. Particle growth curves showing fully reversible hygroscopic growth of
sulfuric acid (H2SO4) particles, deliquescent growth of ammonium
sulfate [(NH4)2 SO4] particles at the deliquescent point (A, about 80%
relative humidity [RH]), reversible hygroscopic growth of ammonium
sulfate solution droplets at RH greater than 80%, and hysteresis (the
droplet remains supersaturated as the RH decreases below 80%) until
the crystallization point (B, about 38% RH) is reached (adapted from
National Research Council, 1993 and Tang, 1980).
Source: National Research Council (1993) adapted from Tang (1980).
1 2.1.3.4 Atmospheric Lifetimes and Removal Processes
2 The lifetimes of particles vary with size. Nuclei-mode particles rapidly grow into the
3 accumulation mode. However, the accumulation mode does not grow into the coarse mode.
4 Accumulation-mode fine particles are kept suspended by normal air motions and have very low
5 deposition rates to surfaces. They can be transported thousands of km and remain in the
6 atmosphere for a number of days. Coarse particles can settle rapidly from the atmosphere within
April 2002
2-25
DRAFT-DO NOT QUOTE OR CITE
-------
216
I I
Theoretical Prediction at 22°c
RH=99.8%
o o o o o Experimental Measurements
RH=99.7%
>___^_——
RH=99.5%
0
50 100 150
NH4 HSO4 Dry Particle Diameter (nm)
200
Figure 2-9. Theoretical predictions and experimental measurements of growth of
NH4HSO4 particles at relative humidity between 95 and 100%.
Source: Lietal. (1992).
1 hours and normally travel only short distances. However, when mixed high into the atmosphere,
2 as in dust storms, the smaller-sized coarse-mode particles may have longer lives and travel
3 greater distances. Dry deposition rates are expressed in terms of a deposition velocity that varies
4 with particle size, reaching a minimum between 0.1 and 1.0 //m aerodynamic diameter (e.g., Lin
5 et al., 1994). Accumulation-mode particles are removed from the atmosphere primarily by cloud
6 processes. Fine particles, especially particles with a hygroscopic component, grow as the relative
7 humidity increases, serve as cloud condensation nuclei, and grow into cloud droplets. If the
8 cloud droplets grow large enough to form rain, the particles are removed in the rain. Falling rain
April 2002
2-26
DRAFT-DO NOT QUOTE OR CITE
-------
1 drops impact coarse particles and remove them. Ultrafine or nuclei-mode particles are small
2 enough to diffuse to the falling drop, be captured, and be removed in rain. Falling rain drops,
3 however, are not nearly as effective in removing accumulation-mode particles as the cloud
4 processes mentioned above. A more detailed discussion of particle deposition, including acid
5 deposition, especially as it applies to deposition to vegetation, soil, and water surfaces, is given in
6 Chapter 4, Environmental Effects of Particulate Matter. Acid deposition and PM are intimately
7 related, first, because particles contribute to the acidification of rain and, secondly, because the
8 gas phase species that lead to dry deposition of acidity are also precursors of particles. Therefore,
9 reductions in SO2 and NOX emissions will decrease both acidic deposition and PM
10 concentrations.
11 Sulfate, nitrate, and some partially oxidized organic compounds are hygroscopic and act as
12 nuclei for the formation of cloud droplets. These droplets serve as chemical reactors in which
13 (even slightly) soluble gases can dissolve and react. Thus, SO2 can dissolve in cloud droplets and
14 be oxidized to sulfuric acid by dissolved ozone or hydrogen peroxide. These reactions take place
15 only in aqueous solution, not in the gas phase. Sulfur dioxide also may be oxidized by dissolved
16 oxygen. This process will be faster if metal catalysts such as iron or manganese are present in
17 solution. If the droplets evaporate, larger particles are left behind. If the droplets grow large
18 enough, they will fall as rain; and the particles will be removed from the atmosphere with
19 potential effects on the materials, plants, or soil on which the rain falls. (Similar considerations
20 apply to dew.) Atmospheric particles that nucleate cloud droplets also may contain other soluble
21 or nonsoluble materials such as metal salts and organic compounds that may add to the toxicity
22 of the rain. Thus, the adverse effects of acid deposition on soils, plants, and trees as well as
23 lakes, streams, and fish may be taken into account in setting secondary PM standards. Sulfuric
24 acid, ammonium nitrate, ammonium sulfates, and organic particles also are deposited on surfaces
25 by dry deposition. The utilization of ammonium by plants leads to the production of acidity.
26 Therefore, dry deposition of particles can also contribute to the ecological damages caused by
27 acid deposition. These effects are discussed in Chapter 4, Environmental Effects of Paniculate
28 Matter.
29
April 2002 2-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.1.4 Summary
2 The physical and chemical properties of ultrafine mode, accumulation mode, and coarse
3 mode particles are summarized in Table 2-1.
4
5
6 2.2 MEASUREMENT OF PARTICULATE MATTER
7 The 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a) summarized
8 sampling and analytical techniques for PM and acid deposition that had appeared in the literature
9 since the earlier 1982 PM AQCD (U.S. Environmental Protection Agency, 1982). Excellent
10 reviews have also been published by Chow (1995) and McMurry (2000). This section discusses
11 problems in measuring PM; new techniques that attempt to alleviate these problems or measure
12 problem species; Federal Reference Methods, speciation monitors, analytical methods for
13 inorganic elements, organic and elemental carbon, and ionic species; and continuous and
14 multiday monitors.
15
16 2.2.1 Particle Measurements of Interest
17 There are many PM components and parameters that are of interest across the various types
18 of uses to which PM measurement data are applied. These uses include analyses of compliance
19 with air quality standards and trends; source category apportionment studies, related to the
20 develop of pollution reduction strategies and the validation of air quality models; studies related
21 to health, ecological, and radiative effects; and characterization of current air quality for
22 presentation to the public in the context of EPA's Air Quality Index. PM measurement
23 components and parameters of specific interest for these various purposes are noted below and
24 summarized in Table 2-2.
25 Particle measurements are needed to determine if a location is in compliance with air
26 quality standards and to determine long-term trends in air quality patterns. For these purposes,
27 precision of the measurements by a variety of measurement instruments in use is a critical
28 consideration. Therefore, intercomparisons of various samplers, under a variety of atmospheric
29 and air quality conditions, are important.
30
April 2002 2-28 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2-1. COMPARISON OF AMBIENT PARTICLES,
FINE MODE (Nuclei Mode Plus Accumulation Mode) AND COARSE MODE
Fine
Coarse
Nuclei
Accumulation
Formed
from:
Formed by:
Composed
of:
Solubility:
Sources:
Combustion, high-temperature
processes, and atmospheric reactions
Nucleation
Condensation
Coagulation
Sulfates
Elemental Carbon
Metal compounds
Organic compounds
with very low
saturation vapor
pressure at ambient
temperature
Probably less soluble
than accumulation
mode
Combustion
Atmospheric
transformation of SO2
and some organic
compounds
High temperature
processes
Atmospheric Minutes to hours
half-life:
Removal
Processes:
Travel
distance:
Grows into
accumulation mode
-------
TABLE 2-2. PARTICULATE MATTER COMPONENTS/PARAMETERS
OF INTEREST FOR HEALTH, ECOLOGICAL, OR RADIATIVE EFFECTS;
FOR SOURCE CATEGORY APPORTIONMENT STUDIES;
OR FOR AIR QUALITY MODELING EVALUATION STUDIES
• Particle number
• Particle surface area
• Particle size distribution
• PM mass (fine PM mass [PM2 5] and coarse thoracic PM mass [PM10_2 5]) including both
nonvolatile mass as measured by the current Federal Reference method and total mass
(including semivolatile components such as ammonium nitrate and semivolatile organic
compounds, but not particle-bound water)
• Ions (sulfate, nitrate, and ammonium)
• Strong acidity (FT)
• Elemental carbon
• Organic carbon (total, nonvolatile, and semivolatile; functional groups and individual
species)
• Transition metals (water soluble, bioavailable, oxidant generation)
• Specific toxic elements and organic compounds
• Crustal elements
• Bioaerosols
• Particle refractive index (real and imaginary)
• Particle density
* Particle size change with changes in relative humidity
1 In order to reduce pollution to attain a standard, local agencies and national research
2 organizations need measurements to identify source categories and to develop and validate air
3 quality models. For these purposes, PM parameters other than mass, such as chemical
4 composition and size distribution, must also be measured. Moreover, measurements are needed
5 with shorter time resolution in order to match changes in pollution with diurnal changes in the
6 boundary layer.
April 2002 2-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 A number of PM measurements are needed for use in epidemiological and exposure studies
2 and to determine components of PM to guide planning and interpretation of toxicologic studies.
3 For these purposes, size and chemical composition measurements are important, as is
4 measurement across different time intervals. For epidemiologic studies of acute (i.e., short-term)
5 PM exposures, 1-h or continuous measurements can provide important information beyond that
6 provided by 24-h measurements. However, for epidemiologic studies of chronic PM exposures,
7 measurements that integrate over longer intervals (e.g., a week to a month) are more relevant.
8 For dosimetric studies and modeling, information will be needed on the particle size distribution
9 and on the behavior of particles as the relative humidity and temperature are increased to those
10 found in the respiratory system.
11 For studies of ecological effects and materials damage, measurements of particles and of
12 the chemical components of particulate matter in rain, fog, and dew are needed (a) to understand
13 the contributions of PM to soiling of surfaces and damage to materials and (b) to understand the
14 wet and dry deposition of acidity and toxic substances to surface water, soil, and plants. Some
15 differentiation into particles size is needed to determine dry deposition.
16 For studies of visibility impairment and radiative effects, information is needed that relates
17 to how particles scatter and absorb light, including refractive index, ratio of scattering to
18 absorption, size distribution, and change in particle size with change in relative humidity.
19 EPA's Air Quality Index is intended to provide the public with near real-time information
20 of air quality in urban areas. For this purpose, PM measurements over short time intervals (e.g.,
21 1-h) or continuous measurements are critical.
22
23 2.2.2 Issues in Measurement of Particulate Matter
24 The EPA decision to revise the PM standards by adding daily and yearly standards for
25 PM2 5 has led to a renewed interest in the measurement of atmospheric particles and also to a
26 better understanding of the problems in developing precise and accurate measurements of
27 particles. It is very difficult to measure and characterize particles suspended in the atmosphere;
28 however, improvements in PM monitoring may be anticipated. EPA's PM standards are based,
29 in part, on epidemiologic relationships between health effects and PM concentrations as
30 measured with existing monitoring methods. As understanding of suspended PM has advanced
31 and new monitoring information has become available, EPA has changed the indicator for the
April 2002 2-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 PM NAAQS from TSP to PM10 to PM2 5. During the current review consideration will be given
2 to a standard for coarse PM.
3 The U.S. Federal Reference Methods (FRM) for PM25 and PM10 provide relatively precise
4 (±10 %) methods for determining the mass of material remaining on a Teflon filter after
5 equilibration. However, numerous uncertainties remain as to the relationship between the mass
6 and composition of material remaining on the filter, as measured by the FRMs, and the mass and
7 composition of material that existed in the atmosphere as suspended PM. As a result, EPA
8 defines accuracy for PM measurements in terms of agreement of a candidate sampler with a
9 reference sampler. Therefore, intercomparisons of samplers become very important in
10 determining how well various samplers agree and how various design choices influence what is
11 actually measured.
12 There are six general areas where choices are made in the design and use of an aerosol
13 sampler. These include (1) treatment of semivolatile components; (2) selection of an upper cut
14 point; (3) separation of fine-mode and coarse-mode PM; (4) treatment of pressure, temperature,
15 and relative humidity; (5) time resolution; and (6) assessment of the reliability of the
16 measurement technique. In many cases, choices have been made without adequate knowledge or
17 understanding of the consequences. As a result, measurement methods developed by different
18 organizations may give different results when sampling the same atmosphere even though the
19 techniques appear to be similar.
20
21 2.2.2.1 Treatment of Semivolatile Components of Particulate Matter
22 Current filtration-based mass measurements can experience significant evaporative losses,
23 during and possibly after collection, of a variety of semivolatile components (i.e., species that
24 exist in the atmosphere in dynamic equilibrium between the condensed phase and gas phase).
25 Important examples include ammonium nitrate, semivolatile organic compounds, and particle-
26 bound water. This problem is illustrated in Figure 2-10.
27 Possible approaches that have been used to address the problem of potentially lost
28 semivolatile components include those that follow, which will be discussed in more detail in
29 subsequent sections.
30 1. Collect/measure all components present in the atmosphere in the condensed phase except
31 particle-bound water. (Examples: Brigham Young absorptive sampler and Harvard pressure
April 2002 2-32 DRAFT-DO NOT QUOTE OR CITE
-------
Should be
retained
Particle-bound water
should be removed
2.5
Aerodynamic Diameter (pm)
i:: i Semivolatile components subject to evaporation during or after sampling
Figure 2-10. Schematic showing major nonvolatile and semivolatile components of PM25.
Semivolatile components are subject to partial to complete loss during
equilibration or heating. The optimal technique would be to remove all
particle-bound water but no ammonium nitrate or semivolatile organic PM.
1 drop monitor. Both require preconcentration of the accumulation mode and reduction of
2 ambient humidity.)
3 2. Stabilize PM at a specified temperature high enough to remove all, or almost all, particle-
4 bound water. This results in loss of much of the semivolatile PM. (Examples: tapered
5 element oscillating microbalance (TEOM) operated at 50 °C, beta gauge with heated inlet.)
6 3. Equilibrate collected material at fixed, near-room temperature and moderate relative humidity
7 to reduce particle-bound water. Accept the loss of an unknown but possibly significant
8 fraction of semivolatile PM. (Example: U.S. Federal Reference Method and most filter-
9 weighing techniques.) Equilibration originally was designed to remove adsorbed water vapor
April 2002
2-33
DRAFT-DO NOT QUOTE OR CITE
-------
1 from glass fiber filters in order to maintain a stable filter weight. The designated RH (40%)
2 was a compromise. If the RH is too low, electrostatic charging becomes a problem. The
3 equilibration process does help provide a stable and reproducible mass. It also reduces the
4 particle-bound water. However, it may not remove all particle-bound water.
5 The amount of semivolatile material lost is dependent on the concentration and
6 composition of the semivolatile components and is, therefore, also dependent on season and
7 location. The amount of semivolatile material lost has been found to be significant in air sheds
8 with high nitrate, wood smoke, or secondary organic aerosols.
9
10 2.2.2.2 Upper Cut Point
11 The upper cut point of the high volume sampler varied with wind speed and direction.
12 Newer PM samplers are usually designed to have an upper cut point and its standard deviation
13 that are independent of wind speed and direction. Current PM samplers have upper cut points
14 that are stable under normal operating conditions. However, problems may occur under unusual
15 or adverse conditions. Ono et al. (2000) recently reported the results of a study in which several
16 PM10 samplers were collocated and operated at various sites at Owens Lake, CA, a location with
17 high concentrations of coarse PM. Samplers included the Partisol sampler, the TEOM, a
18 dichotomous sampler, the Wedding high-volume sampler, and the Graseby high-volume sampler.
19 They found that the TEOM and Partisol samplers agreed to within 6% on average. The
20 dichotomous sampler and the Graseby and Wedding high-volume samplers, however, measured
21 significantly lower PM10 concentrations than the TEOM (on average 10, 25, and 35% lower,
22 respectively). These lower concentrations were attributed to a decrease in cut point at higher
23 wind speeds and possibly when the inlet is dirty.
24 The choice of the cut point characteristics depends upon the application for the sampling
25 device. A separation that simulates the removal of particles by the upper part of the human
26 respiratory system might appear to be a good choice for both health risk and regulatory
27 monitoring (i.e., measure what gets into the lungs). The ACGIH-ISO-CEN penetration curve for
28 thoracic particles (particles able to pass the larynx and penetrate into the bronchial and alveolar
29 regions of the lung) has a 50% cut point at 10 //m aerodynamic diameter (Da). The U.S. PM10
30 separation curve is sharper than the thoracic penetration curve but has the advantage of reducing
April 2002 2-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 the problem of maintaining the finite collection efficiency specified by the thoracic penetration
2 curve for particles larger than 10 //m Da. (See Section 2.1.2.2 and Figure 2-6).
3
4 2.2.2.3 Cut Point for Separation of Fine-Mode and Coarse-Mode Particulate Matter
5 As Table 2-1 showed, fine- and coarse-mode particles differ not only in size and
6 morphology (e.g., smooth droplets versus rough solid particles), but also in formation
7 mechanisms; sources; and chemical, physical, and biological properties. They also differ in
8 concentration-exposure relationships, dosimetry (deposition in the respiratory system), toxicity,
9 and health effects as observed by epidemiologic studies. Thus, it is desirable to measure fme-
10 mode PM and coarse-mode PM separately in order to properly allocate health effects to either
11 fine-mode or coarse-mode PM and to correctly determine sources by receptor modeling
12 approaches. For example, sulfate in the fine-mode is associated with hydrogen or ammonium
13 ions while sulfate in the coarse mode is associated with basic metal ions. Transition metals in
14 the coarse mode are likely to be associated with soil and tend to be less soluble (and presumably
15 less bioavailable) than transition metals in fresh combustion particles found in the fine mode.
16 The 2.5 //m Da cut point was chosen in the early 1970s as the cut point for a new
17 dichotomous sampler (Loo et al., 1976; Jaklevic et al., 1977) for use in the Regional Air
18 Pollution Study in St. Louis, MO. At that time aerosol scientists were beginning to realize that
19 there was a minimum between 1 and 3 //m in the distribution of particle size by volume (Whitby
20 et al., 1974). The 2.5 //m cut point was subsequently used as an indicator of fine-mode PM in a
21 number of studies, including the Harvard Six-City Studies of the relationships between mortality
22 and PM concentrations (Dockery et al., 1993; Schwartz et al., 1996). A 2.5 //m cut point was
23 also used in the Inhalable Particle Network (Suggs and Burton, 1983) which provided data for
24 another major epidemiologic study of PM - mortality relationships using an American Cancer
25 Society cohort (Pope et al., 1995). Therefore, at the time of the last review of the NAAQS for
26 PM (U.S. Environmental Protection Agency, 1996a,b), there were a number of epidemiologic
27 studies demonstrating a statistical relationship between PM2 5 concentrations and mortality.
28 It is now understood that the intermodal region (1-2.5 //m) may contain either
29 accumulation-mode or coarse-mode material and that the two modes may overlap in this region.
30 The experimental information on the composition and source of the intermodal mass was
31 discussed extensively in the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a).
April 2002 2-3 5 DRAFT-DO NOT QUOTE OR CITE
-------
1 Depending on conditions, a significant amount of either accumulation- or coarse-mode material
2 may be found in the intermodal region between 1 and 2.5 //m. The analysis demonstrated the
3 important role of relative humidity in influencing the size of particles in both the accumulation
4 and coarse modes.
5 As the relative humidity increases, hygroscopic accumulation-mode particles will increase
6 in size due to accumulation of particle-bound water. At high relative humidities, some originally
7 submicrometer accumulation-mode PM may be found with a Da above 1 //m. At a relative
8 humidity of 100%, such as found in fog and clouds, accumulation mode PM may exceed 2.5 //m
9 Da. What is not well understood is whether such particles will shrink to diameters below 1 //m as
10 the RH decreases or whether reactions occurring in the wet particles will result in an increase in
11 non-aqueous mass so that even at low RH the diameters would exceed 1 //m. On the other hand,
12 at very low relative humidity, coarse-mode particles may be fragmented into smaller sizes, and
13 small amounts of coarse-mode PM may be found with an Da below 2.5 //m (Lundgren et al.,
14 1984; Lundgren and Burton, 1995). Thus, a PM25 sample will contain all of the fine-mode
15 material except during periods of RH near 100 %. However, under conditions of low RH, it may
16 also contain a small fraction of the coarse-mode PM. The selection of a cut point of 2.5 //m as a
17 basis for EPA's 1997 NAAQS for fine particles (Federal Register, 1997) and its continued use in
18 many health effects studies reflect the importance placed on more complete inclusion of fme-
19 mode particles, while recognizing that intrusion of coarse-mode particles can occur under some
20 conditions with this cut point.
21 In addition to the influence of relative humidity, in areas where winds cause high
22 concentrations of windblown soil, there is evidence that a significant amount of coarse-mode PM
23 may be found below 2.5 //m. An example, taken from data collected during the August 1996
24 dust storm in Spokane, WA, is shown in Figure 2-11. Note that the PM10 scale is 10 times that of
25 the other size fractions. PMl3 although high in the morning, goes down as the wind increases and
26 PM10, PM25, and PM^.j go up. During the peak of the dust storm, PM^.j was 88% of PM25.
27 For the 24-h period, PM2 54 was 54% of PM2 5. However, PMl was not affected by the intrusion
28 of coarse-mode particles. Similar considerations probably apply to intrusions of dust transported
29 from distance sources such as the Sahara and Gobi deserts (Husar et al., 2001).
30 A cut point of 1 //m could reduce the misclassification of coarse-mode material as fine,
31 especially in a areas with high levels of wind blown soil, but under high RH conditions could
April 2002 2-36 DRAFT-DO NOT QUOTE OR CITE
-------
-IIIIIIIIIIIIIIIIIIIII
12am 2am 4am 6am Sam 10am 12pm 2pm 4pm 6pm 8pm 10pm
Local Time, August 30,1996
Figure 2-11. Particulate matter concentrations in Spokane, WA, during the August 30,
1996 dust storm.
Source: Claiborn et al. (2000).
1 result in some fine-mode material being misclassified as coarse. A reduction in RH, either
2 intentionally or inadvertently, will reduce the size of the fine mode. A sufficient reduction in RH
3 should yield a dry fine-particle mode with very little material above 1.0 //m. Studies of the
4 changes in particle size with changes in relative humidity suggest that only a small fraction of
5 accumulation mode particles will be above 1 //m in diameter at RH below 60%, but a substantial
6 fraction will grow above 1 //m for RH above 80% (Hitzenberger et al., 1997; McMurry and
7 Stolzenburg, 1989; U.S. Environmental Protection Agency, 1996a).
8 Under high relative humidity circumstances, a monitor using a 1.0 //m Da cut point can
9 achieve better modal separation if the air stream is dehumidified to some fixed humidity that
10 would remove all or most particle-bound water without evaporating semivolatile components.
11 New techniques which require reduction of RH prior to collection have been developed for
12 measurement of fine particulate matter minus particle-bound water but including semivolatile
13 nitrate and organic compounds. With such techniques, PMl measurements, in conjunction with
14 concurrent PM2 5 measurements, would be useful for exposure, epidemiologic, and source
April 2002
2-37
DRAFT-DO NOT QUOTE OR CITE
-------
1 apportionment studies, especially in areas where intrusion of coarse-mode particles into the
2 intermodal range is likely.
3
4 2.2.2.4 Treatment of Pressure, Temperature, and Relative Humidity
5 There are a variety of techniques for defining (or ignoring) the pressure, temperature, and
6 relative humidity during and after sampling. For example, the sample volume may be based on
7 the mass or volumetric flow corrected to standard temperature and pressure (273 °K and 1 atm.)
8 (current EPA technique for PM10), or it may be on the volumetric flow at ambient conditions of
9 temperature and pressure (current EPA technique for PM2 5).
10 There are also a variety of options for the control of temperature during collection. The
11 particles may be heated enough to remove much of the particle-bound water (i.e., TEOM at
12 50 °C); the particles may be heated several degrees, just enough to prevent condensation of water
13 in sampling system; the particles and the sampler may be maintained near ambient temperature
14 (±5 °C of ambient temperature is required for EPA FRMs); or the particles and sampler may be
15 maintained at constant temperature inside a heated or air conditioned shelter. There are also
16 options for control of temperature after collection: (a) no control (room temperature) or (b) ship
17 and store at cool temperature (4 °C is the current EPA FRM requirement).
18 Consideration must also be given to relative humidity. Changes in relative humidity cause
19 changes in particle size of hygroscopic or deliquescent particles. Changing relative humidity by
20 adding or removing water vapor affects measurements of particle number, particle surface area,
21 and particle size distribution and the amount of overlap of fine-mode and coarse-mode particles.
22 Changing relative humidity by intentional or inadvertent changes in temperature also affects the
23 amount of loss of ammonium nitrate and semivolatile organic compounds. Monitoring personnel
24 should be aware of the various options for treatment of pressure, temperature, and relative
25 humidity; make appropriate selections; and document which options are used.
26 Studies of relationships between personal/indoor/outdoor measurements present special
27 problems. Indoor environments are typically dryer than outdoors and may be warmer or, if
28 air-conditioned, cooler. These differences may change particle size and the amount of
29 volatilization of semivolatile components. Such changes between indoors and outdoors will
30 complicate the comparison of indoor to outdoor PM concentrations; the modeling of personal
April 2002 2-3 8 DRAFT-DO NOT QUOTE OR CITE
-------
1 exposure to all particles; and apportionment of personal exposure into particles of ambient
2 origin, particles of indoor origin, and personal activity particles.
3
4 2.2.2.5 Time Resolution
5 The classic 24-hour filter collection technique is being supplemented by a variety of
6 continuous monitors for various PM parameters. This process is being accelerated by the lower
7 operational cost of continuous monitors and the availability of new continuous monitors for
8 mass, number, and certain chemical components, as well as older methods based on beta
9 attenuation or light scattering. Most epidemiologic studies have used 24-hour concentrations as
10 exposure indicators. However, one epidemiologic study of chronic effects uses a filter sampler
11 with a two-week collection period (Gauderman et al., 2000). Another recent study used 1-2 h
12 concentrations (see Peters et al., 2000). Continuous methods are discussed in Section 2.2.5.
13
14 2.2.2.6 Accuracy and Precision
15 Precision is typically determined by comparison of collocated samplers or through replicate
16 analyses; whereas accuracy is determined through the use of traceable calibration standards.
17 Unfortunately, no standard reference calibration material or procedure has been developed for
18 suspended, atmospheric PM. It is possible to determine the accuracy of certain components of
19 the PM measurement system (e.g., flow control, inlet aspiration, PM25 cut, weighing, etc.). The
20 absolute accuracy for collecting a test aerosol can also be determined by isokinetic sampling in a
21 wind tunnel. However, it is not currently feasible to provide a simulated atmospheric aerosol
22 with naturally occurring semivolatile components. It is particularly challenging to develop an
23 atmospheric aerosol calibration standard suitable for testing samplers in the field. Therefore, it is
24 not possible at the present time to establish the absolute accuracy of a PM monitoring technique.
25 Intercomparison studies may be used to establish the precision of identical monitors and the
26 extent of agreement between different types of monitors. Such studies are important for
27 establishing the reliability of PM measurements. Intercomparison studies have contributed
28 greatly to our understanding of the problems in PM measurement. Such studies will be discussed
29 as they apply to specific measurement problems, monitoring instruments, or analytical
30 techniques.
April 2002 2-39 DRAFT-DO NOT QUOTE OR CITE
-------
1 Some measurement errors of concern in PM10 sampling, including those that arise due to
2 uncertainty tolerances in cutpoint, particle bounce and reentrainment, impactor surface
3 overloading, and losses to sampler internal surfaces, were discussed in detail in the 1996 PM
4 AQCD (U.S. Environmental Protection Agency, 1996a). Other measurement errors of concern in
5 PM25 sampling arise not only because of our inability to assess accuracy in an absolute sense due
6 to a lack of an atmospheric aerosol calibration standard, but also because of the inclusion in
7 PM2 5 of a small amount of coarse particles as discussed in Section 2.2.1.3 and because of
8 problems associated with the definition of PM25 as what remains after collection on a filter and
9 equilibration rather than the mass of particles as they exist in the air. Still, it is possible to
10 measure PM indicators with high precision. Detailed information on precision and quality
11 assurance may be found on EPA's Technology Transfer Network website (U.S. Environmental
12 Protection Agency, 2002). See discussion in Section 2.2.4.
13 Because of the difficulties associated with determining the accuracy of PM measurements,
14 EPA has sought to make FRM measurements equivalent by specifying operating conditions and,
15 in the case of PM25 samplers, by specifying details of the sampler design. Thus, both the PM10 as
16 well as the PM25 standards are defined with consistency of measurement technique rather than
17 with the accuracy of the true mass concentration measurement in mind (McMurry, 2000). It is
18 acknowledged in the Federal Register (1997) that, "because the size and volatility of the particles
19 making up ambient particulate matter vary over a wide range and the mass concentration of
20 particles varies with particle size, it is difficult to define the accuracy of PM25 measurements in
21 an absolute sense...." Thus, accuracy is defined as the degree of agreement between a field PM2 5
22 sampler and a collocated PM2 5 reference method audit sampler (McMurry, 2000). The Federal
23 Reference Method for PM25 is discussed in Section 2.2.3.3. As mentioned earlier, volatilization
24 of organic compounds and ammonium nitrate during sampling or post-sampling handling can
25 lead to significant underestimation of the fine parti culate mass concentration in some locations.
26 Sources of error in the measurement of mass of PM25 suspended in the atmosphere also arise
27 because of adsorption or desorption of semivolatile vapors onto or from collected PM, filter
28 media, or other sampler surfaces; neutralization of acid or basic vapors on either filter media or
29 collected PM; and artifacts associated with particle-bound water.
30 During the past 25 years, there have been advancements in the generation and classification
31 of monodisperse aerosols, as well as in the development of electron microscopy and imaging
April 2002 2-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 analysis, that have contributed to the advancement in aerosol calibration (Chen, 1993). Still, one
2 of the limitations in PM sampling and analysis remains the lack of primary calibration standards
3 for evaluating analytical methods and for intercomparing laboratories. Klouda et al. (1996)
4 examined the possibility of resuspending the National Institute of Science and Technology
5 (NIST) Standard Reference Material 1649 (Urban Dust) in air for collection on up to 320 filters
6 simultaneously using Standard Research International's dust generation and collection system.
7 However, the fine component is not resuspended and the semivolatile component has evaporated.
8 Consequently, this material is not a suitable standard for suspended PM. Little additional work
9 in this area has been reported.
10 Methods validation was discussed in the 1996 PM AQCD (U.S. Environmental Protection
11 Agency, 1996a), and the usefulness of intercomparisons and "internal redundancy" was
12 emphasized. For example, a number of internal consistency checks are applied to the IMPROVE
13 network (Malm et al., 1994). These include mass balances, sulfur measurements by both proton
14 induced X-ray emission (PIXE) and ion chromatography (1C), and comparison of organic matter
15 by combustion and by proton elastic scattering analysis (PESA) analysis of hydrogen. Mass
16 balances compare the gravimetrically determined mass with the mass calculated from the sum of
17 the major chemical components (i.e., crustal elements plus associated oxygen, organic carbon,
18 elemental carbon, sulfate, nitrate, ammonium, and hydrogen ions). Mass balances are useful
19 validation techniques; however, they do not check for, or account for, artifacts associated with
20 the absorption of gases during sampling or the loss of semivolatile material during sampling.
21 The mass balance check may appear reasonable even if such artifacts are present because only the
22 material collected on the filter is included in the balance.
23
24 2.2.3 Measurement of Semivolatile Particulate Matter
25 It is becoming increasingly apparent that the semivolatile component of PM may
26 significantly affect the quality of the measurement and can lead to both positive and negative
27 sampling artifacts. Loss of semivolatile species, like ammonium nitrate and many organic
28 species, may occur during sampling because of changes in temperature, relative humidity, or
29 composition of the aerosol or because of the pressure drop across the filter (McMurry, 2000).
30 Gas phase organic species, both volatile and semivolatile, may adsorb onto or react with filter
31 media or collected PM, leading to a positive sampling artifact. Quartz fiber filters have a large
April 2002 2-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 specific surface area on which adsorption of gases can occur. A number of other types of filters
2 (e.g., stretched Teflon membrane filters) have much smaller exposed surface areas (Turpin et al.,
3 1994) and appear to be subject to less adsorption (Kirchstetter et al., 2001; Turpin et al., 1994).
4 Tsai and Huang (1995) observed positive sulfate and nitrate artifacts on high-volume PM10 quartz
5 filters and attributed the artifacts to interactions between acidic gases (SO2, HONO, and HNO3)
6 and both the filter media (either glass fiber or quartz) and the coarse particles collected on the
7 filter. Volatilization losses also have been reported to occur during sample transport and storage
8 (Chow, 1995). Evaporative losses of particulate nitrates have been investigated in laboratory and
9 field experiments (e.g., Wang and John, 1988) and in theoretical studies (Zhang and McMurry,
10 1992). The results of recent studies that focused on volatilization losses of parti culate nitrates are
11 discussed in more detail in Section 2.2.3.1.
12 The theory describing phase equilibria of semivolative organic compounds (SVOC)
13 continues to be developed. Liang et al. (1997), Jang et al. (1997), and Strommen and Kamens
14 (1997) have modeled the gas/particle partitioning of SVOC on inorganic, organic, and ambient
15 smog aerosols.
16 The positive artifact associated with adsorption of organic vapors onto quartz filters has
17 been examined in experiments in which two quartz fiber filters were deployed in series. The
18 second quartz filter may indicate gaseous volatile organic compounds (VOC) adsorbed on both
19 filters (positive artifact), SVOC evaporated from particles on the first filter and subsequently
20 adsorbed on the second filter (negative artifact), or a combination of both effects. Unless the
21 individual compounds are identified, the investigator does not know what to do with the loading
22 value on the second filter (i.e., to add or subtract from the first filter loading value). Moreover,
23 even if the individual compounds were identified on the back-up filter, the decision concerning
24 adding or subtracting the back-up filter loading would not be straightforward.
25 The developing state of the art in which diffusion denuder technology is being applied to
26 SVOC sampling (e.g., Eatough et al., 1993; Gundel et al., 1995), as well as for sampling of gas
27 and parti culate phase organic acids (Lawrence and Koutrakis, 1996a,b), holds promise for
28 improving the understanding of SVOC sampling artifacts. In a denuder-based system, gas-phase
29 organics are removed by diffusion to an adsorbent surface (e.g., activated carbon, special
30 polymer resins, etc.). Particles then are collected on a filter downstream of the denuder and the
31 remaining organic vapors (i.e., from denuder breakthrough and volatile losses from the collected
April 2002 2-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 particles) are collected in an adsorbent downstream of the filter (e.g., charcoal or carbon-
2 impregnated filters, polyurethane foam, or polystyrene-divinylbenzene resin [XAD]). The results
3 of recent studies that have focused on treatment of both positive and negative sampling artifacts
4 associated with SVOC are discussed in Section 2.2.3.2. Detailed information on the use of
5 denuder systems to measure semivolatile species is provided in Section 2.2.3.3.
6 Finally, Eatough et al. (1999a) have reported on a batch sampler that attempts to correct
7 simultaneously for volatilization losses of both nitrate and SVOC. These samplers are also
8 discussed in more detail in Section 2.2.3.2.
9
10 2.2.3.1 Particulate Nitrates
11 It is well known that volatilization losses of particulate nitrates (e.g., Zhang and McMurry
12 [1992]; see also Hering and Cass [1999] and references therein) occur during sampling on Teflon
13 filters. The affect on the accuracy of atmospheric particulate measurements from these
14 volatilization losses is more significant for PM2 5 than for PM10. The FRM for PM2 5 will likely
15 suffer loss of nitrates similar to that experienced with other simple filter collection systems.
16 Sampling artifacts resulting from the loss of particulate nitrates represents a significant problem
17 in areas such as southern California that experience high amounts of nitrates. Hering and Cass
18 (1999) examined the errors in PM25 mass measurements because of volatilization of particulate
19 nitrate by looking at data from two field measurement campaigns conducted in southern
20 California: (1) the Southern California Air Quality Study (SCAQS) (Lawson, 1990) and (2) the
21 1986 CalTech study (Solomon et al., 1992). In both these studies, side-by-side sampling of PM2 5
22 was conducted. One sampler collected particles directly onto a Teflon filter. The second
23 sampler consisted of a denuder to remove gaseous nitric acid followed by a nylon filter that
24 absorbs the HNO3 which evaporates from ammonium nitrate. In both studies, the denuder
25 consisted of MgO-coated glass tubes (Appel et al., 1981). Fine particulate nitrate collected on
26 the Teflon filter was compared to fine particulate nitrate collected on the denuded nylon filter.
27 In both studies, the PM2 5 mass lost because of volatilization of ammonium nitrate represented a
28 significant fraction of the total PM2 5 mass. The fraction of mass lost was higher during summer
29 than during fall (17% versus 9% during the SCAQS study and 21% versus 13% during the
30 CalTech study; Figure 2-12). In regard to percentage loss of nitrate, as opposed to percentage
April 2002 2-43 DRAFT-DO NOT QUOTE OR CITE
-------
0)
E
'>
ro
O
•5
80%
60% -
40% -
SCAQS Data Set
o Summer Measurements
• Fall Measurements
20% -
50
100
150
200
250
PM2.5 Gravimetric Mass (ng/m )
80%
-------
1 Finally, during the SCAQS study, particulate samples also were collected using a Berner
2 impactor and greased Tedlar substrates in size ranges from 0.05 to 10 //m in aerodynamic
3 diameter. The Berner impactor PM2 5 nitrate values were much closer to those from the denuded
4 nylon filter than those from the Teflon filter with the impactor nitrate being approximately
5 2% lower than the nylon filter nitrate for the fall measurements and approximately 7% lower
6 during the summer measurements. When the impactor collection was compared to the Teflon
7 filter collection for a nonvolatile species (sulfate), the results were in agreement.
8 It should be noted that filters or collection surfaces were removed immediately after
9 sampling and placed into vials containing a basic extraction solution during these
10 intercomparison studies. Therefore, losses that might occur during handling, storage, and
11 equilibration of filters or impaction surfaces were avoided. The loss of nitrate observed from
12 Teflon filters and impaction surfaces in this study, therefore, is a lower limit compared to losses
13 that might occur during the normal processes involved in equilibration and weighing of filters
14 and impaction surfaces. Brook and Dann (1999) observed much higher nitrate losses during a
15 study in which they measured particulate nitrate in Windsor and Hamilton, Ontario, Canada, by
16 three techniques: (1) a single Teflon filter in a dichotomous sampler, (2) the Teflon filter in an
17 annular denuder system (ADS), and (3) total nitrate including both the Teflon filter and the nylon
18 back-up filter from the ADS. The Teflon filter from the dichotomous sampler averaged only
19 13% of the total nitrate. The Teflon filter from the ADS averaged 46% of the total nitrate. The
20 authors concluded that considerable nitrate was lost from the dichotomous sampler filters during
21 handling, which included weighing and x-ray fluorescence (XRF) measurement in a vacuum.
22 Kim et al. (1999) also examined nitrate sampling artifacts by comparing denuded and
23 undenuded quartz and nylon filters, during the PM10 Technical Enhancement Program (PTEP) in
24 the South Coast Air Basin of California. They observed negative nitrate artifacts (losses) for
25 most measurements; however, for a significant number of measurements they observed positive
26 nitrate artifacts. Kim et al. (1999) pointed out that random measurement errors make it difficult
27 to measure true amounts of nitrate loss.
28 Several diffusion denuder samplers have been developed to account for the nitrate lost
29 because of volatilization from filters, many of which were discussed in the 1996 PM AQCD
30 (U.S. Environmental Protection Agency, 1996a). Eatough et al. (1999a) developed a high-
31 volume diffusion denuder system in which diffusion denuder and particle concentrator
April 2002 2-45 DRAFT-DO NOT QUOTE OR CITE
-------
1 techniques were combined (see Section 2.2.3.2). The particle concentrator reduces the flow
2 through the denuder so that the denuder can be operated for weeks without a loss of collection
3 efficiency, making the sampler suitable for routine field sampling. The system was evaluated for
4 the collection of fine particulate sulfate and nitrate in Riverside, CA (Eatough et al., 1999b).
5 Concentrations of PM25 nitrate obtained from the PC-BOSS agreed with those obtained using the
6 Harvard-EPA Annular Denuder Sampler, HEADS (Koutrakis et al., 1988).
7
8 2.2.3.2 Semivolatile Organic Compounds
9 In addition to their contribution to suspended PM mass, SVOC are also of interest because
10 of their possible health effects. SVOC include products of incomplete combustion such as
11 polycyclic aromatic hydrocarbons (PAHs) and polycyclic organic matter, which has been
12 identified as a hazardous air pollutant. PAHs also have been suggested as alternative particulate
13 tracers for automobile emissions because the phase-out of organo-lead additives to gasoline
14 means that lead is no longer a good tracer for automobiles (Venkataraman et al., 1994). PAHs
15 also are emitted during biomass burning, including burning of cereal crop residues and wood
16 fuels (Jenkins et al., 1996; Roberts and Corkill, 1998).
17 The positive quartz filter artifact was previously mentioned and has been discussed by
18 others (Gundel et al., 1995; Turpin et al., 2000). It is also possible that some SVOC may desorb
19 from the filter resulting in a negative artifact (Eatough et al., 1993; Tang et al., 1994; Eatough
20 et al., 1995; Gundel et al., 1995; Cui et al., 1998; Pang et al., 2001; Finn et al., 2001).
21 Semivolatile organic compounds can similarly be lost from Teflon filters because of
22 volatilization, causing the PM2 5 mass to be significantly underestimated (negative artifact). Like
23 particulate nitrates, the FRM for PM2 5 will suffer loss of SVOC, similar to the losses
24 experienced with other simple filter collection systems. Most studies that have focused on the
25 positive and negative sampling artifacts associated with SVOC compounds have utilized either
26 diffusion denuder technology or placed an adsorbent media, such as a back-up quartz filter or a
27 polyurethane foam adsorbent behind the main filter.
28 Using their multichannel diffusion denuder sampling system (BOSS), Eatough et al. (1995)
29 reported that, for samples collected at the South Coast Air Quality Management District
30 sampling site at Azusa, CA, changes in the phase distribution of SVOC could result in a loss on
31 average of 35% of the particulate organic material. Cui et al. (1998) found that losses of SVOC
April 2002 2-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 from particles in the Los Angeles Basin during the summer were greater during the night
2 (average = 62%) than during the day (average = 42%).
3 The percent SVOC lost from the front filter in a filter-denuder system may be greater than
4 that lost in a filter-only system such as the FRM. In a filter-denuder system, the gas-phase
5 component of the SVOC is removed. The absence of the gas-phase causes the gas-particle
6 equilibrium to shift so the SVOC collected on the filter may evaporate more rapidly in a filter-
7 denuder system than in a filter-only collection system. To determine the fraction of SVOC lost
8 from a Teflon filter in a filter-only system, it is necessary to compare the amount measured by a
9 nondenuder system with that measured by a denuder system. At present, little information is
10 available on the volatilization losses of SVOC. However, in one study (Pang et al., 2001), the
11 total mass on denuded and undenuded filters were compared and found to be identical within
12 error limits (R2 = 0.816, slope = 0.961 ± 0.027 for total mass compared to R2 = 0.940, slope =
13 0.986 ± 0.020 for sulfate). Pang et al. interpreted this result as suggesting that the major cause of
14 loss of SVOC is the pressure drop across the filter.
15 Positive artifacts may occur during sample collection because of the adsorption of gases
16 onto the filter materials (e.g., Gundel et al., 1995). Using a quartz filter behind a Teflon filter,
17 Kim et al. (2001) estimated that on an annual average basis 30% of the PM25 organic carbon
18 concentration resulted from positive artifacts. There is a larger positive artifact because of
19 greater adsorption of organic vapor onto quartz fiber filters than onto Teflon filters (Turpin et al.,
20 1994; Chow et al., 1994, 1996; Eatough et al., 1996; Finn et al., 2001).
21 Kirchstetter et al. (2001) report that adsorptive properties of quartz fiber filters vary with lot
22 number; therefore, front and back-up filters should be taken from the same lot. Recent literature
23 suggests that a Teflon filter followed by a quartz back-up filter appears to provide a better
24 estimate of the adsorption of gases on a quartz fiber front filter than does a quartz filter followed
25 by a quartz backup and that the difference between these two adsorption estimates can be
26 substantial for short durations (Novakov et al., 1997; Kirchstetter et al., 2001; Turpin et al.,
27 2000). The typically lower organic carbon loadings on concurrently collected quartz followed by
28 quartz filters relative to Teflon followed by quartz filters are believed to occur because
29 adsorption on the quartz front filter acts to reduce the gas-phase concentration downstream until
30 adsorption equilibrium has been achieved in the vicinity of the front quartz filter surface.
31 Because Teflon filters have little affinity for organic vapors, this equilibrium occurs almost
April 2002 2-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 instantaneously for Teflon filters, and the Teflon-quartz back-up filter is exposed to the ambient
2 concentration of organic vapors from the beginning of the sampling period. It might be expected
3 that the quantity of organic vapor adsorbed on quartz filters would depend on the organic
4 composition and would vary by season and location.
5
6 Use ofDenuder Systems To Measure Semivolatile Compounds
1 Phase distribution of semivolatile organic species has been the subject of several studies
8 that have employed denuder technology (see Gundel et al., 1995; Gundel and Lane, 1999) to
9 directly determine the phase distributions while avoiding some of the positive and negative
10 sampling artifacts associated with using back-up quartz filters. In an ideal system with a denuder
11 that is 100% efficient, the gas phase would be collected in the denuder and the particle phase
12 would be the sum of the material collected on the filter and the adsorbent downstream. Denuder
13 collection efficiency depends on the denuder surface area (+), the diffusivity (+) and vapor
14 pressure (-) of the compound, the temperature (-) and flow rate (-) of the air stream, and the
15 presence of competing species (-), including water vapor (Cui et al., 1998; Kamens and Coe,
16 1997; Lane et al., 1988). (The + and - symbols in parentheses indicate qualitatively the effect
17 increasing each parameter would have on efficiency). In a system with a denuder collection
18 efficiency less than 100%, the collection efficiency must be known to accurately attribute
19 adsorbed organics from denuder breakthrough to the gas phase and adsorbed organics volatilized
20 from collected particles to the particle phase. In calculating the overall phase distributions of
21 SVOC PAH from a denuder system, the collection efficiency for each compound is needed.
22 The efficiency of silicone-grease-coated denuders for the collection of polynuclear aromatic
23 hydrocarbons was examined by Coutant et al. (1992), who examined the effects of uncertainties
24 in the diffusion coefficients and in the collisional reaction efficiencies on the overall phase
25 distributions of SVOC PAH calculated using denuder technology. In their study, they used a
26 single stage, silicone-grease-coated aluminum annular denuder with a filter holder mounted
27 ahead of the denuder and an XAD trap deployed downstream of the denuder. In a series of
28 laboratory experiments, they spiked the filter with a mixture of perdeuterated PAH, swept the
29 system with ultra-high purity air for several hours, and then analyzed the filter and the XAD.
30 They found that the effects of these uncertainties, introduced by using a single compound as a
31 surrogate PAH (in their case, naphthalene) for validation of the denuder collection efficiency, are
April 2002 2-48 DRAFT-DO NOT QUOTE OR CITE
-------
1 less significant than normal variations because of sampling and analytical effects. Results on
2 field studies using their sampling system have not been published.
3 For measuring paniculate phase organic compounds, the denuder-based sampling system
4 represents an improvement over the filter/adsorbent collection method (Turpin et al., 1993).
5 Some researchers, however, have reported that denuder coatings themselves can introduce
6 contamination (Mukerjee et al., 1997) and that the adsorbed species may be difficult to remove
7 from the coating (Eatough et al., 1993).
8 In a study conducted in southern California (Eatough et al., 1995), the Brigham Young
9 University Organic Sampling System (BOSS; Eatough et al., 1993) was used for determining
10 POM composition, and a high-volume version (BIG BOSS; flow rate 200 L/min) was utilized for
11 determining the particulate size distribution and the chemical composition of SVOC in fine
12 particles. The BOSS, a multi-channel diffusion denuder sampling system, consists of two
13 separate samplers (each operating at 35 L/min). The first sampler consists of a multi-parallel
14 plate diffusion denuder with charcoal-impregnated filter papers as the collection surfaces
15 followed by a two-stage quartz filter pack and a two-stage charcoal-impregnated filter pack. The
16 second sampler operating in parallel with the first consists of a two-stage quartz filter pack,
17 followed by the parallel plate denuder, followed by the two-stage charcoal-impregnated filter
18 pack. The filter samples collected by the BOSS sampler were analyzed by temperature-
19 programmed volatilization analysis. The second channel allows calculations of the efficiency of
20 the denuder in removing gas-phase specifics that would be absorbed by the charcoal impregnated
21 filter. Eatough et al. (1995) also operated a two-stage quartz filter pack alongside the BOSS
22 sampler. The BIG BOSS system (Tang et al., 1994) consists of four systems (each with a
23 flowrate of 200 L/min). Particle size cuts of 2.5, 0.8, and 0.4 //m are achieved by virtual
24 impaction, and the sample subsequently flows through a denuder, then is split, with the major
25 flow (150 L/min) flowing through a quartz filter followed by an XAD-II bed. The minor flow is
26 sampled through a quartz filter backed by a charcoal-impregnated filter paper. The samples
27 derived from the major flow (quartz filters and XAD-II traps) were extracted with organic
28 solvents and analyzed by gas chromatography (GC) and GC-mass spectroscopy. The organic
29 material lost from the particles was found to represent all classes of organic compounds.
30 Eatough et al. (1996) operated the BOSS sampler for a year at the IMPROVE site at
31 Canyonlands National Park, UT, alongside the IMPROVE monitor and alongside a separate
April 2002 2-49 DRAFT-DO NOT QUOTE OR CITE
-------
1 sampler consisting of a two-stage quartz filter pack. They found that concentrations of
2 particulate carbon determined from the quartz filter pack sampling system were low on average
3 by 39%, and this was attributed to volatilization losses of SVOC from the quartz filters.
4 In another study conducted with the BOSS in southern California, losses of 35% of the POM, on
5 average, were found and attributed to losses of the SVOC during sampling (Eatough et al., 1995).
6 The denuder used in the various BOSS samplers consists of charcoal-impregnated cellulose
7 fiber filter material. Denuder collection efficiencies of greater than 95% have been reported for
8 organic gases that adsorb on quartz and charcoal-impregnated filters (Eatough et al., 1999a; Ding
9 et al., 2002; Lewtas et al., 2001). However, because the mass concentration of gas phase species
10 that adsorb on quartz and charcoal-impregnated filters is so much greater than the mass of
11 semivolatile organic material in the particulate phase, it is necessary to measure and account for
12 the inefficiency of the denuder in the BOSS samplers. To address this problem, Brigham Young
13 University (BYU) developed a particle-concentrator (PC)-BOSS system (Ding et al., 2002;
14 Eatough et al., 1999b; Lewtas et al., 2001; Modey et al., 2001; Pang et al., 2001, 2002a,b). The
15 PC-BOSS includes a virtual impactor upstream of the denuder to improve the denuder collection
16 efficiency by removing a majority of the gases from the aerosol flow. With this system, denuder
17 collection efficiencies of greater than 99% have been reported for organic gases, SO2(g),
18 HNO3(g) and other species that adsorb on quartz and charcoal-impregnated filters (Pang et al.,
19 2001). Since the concentrations of semivolatile organic and other gases in the presence of the
20 concentrated particles is not altered by this process, the gas-particle distribution should be little
21 altered by the concentration process. The virtual impactor has a 50% cut point at 0.1 //m
22 aerodynamic diameter. As a result, some particles in the 0.05 to 0.2 //m diameter size range will
23 be removed in the major flow along with the majority of the gases. Therefore, the mass
24 collection efficiency of the virtual impactor concentrator will be a function of the particle size
25 distribution in the 0.05 to 0.1 //m size range. This collection efficiency is measured by
26 comparing the concentration of nonvolatile components measured in the concentrated sample
27 with that measured in an unconcentrated sample. The concentration efficiency varies from 50 to
28 75%. It is relatively constant over periods of weeks but varies by season and site, presumably as
29 the particle size distribution changes. Previous studies at Harvard (Sioutas et al., 1995a,b) have
30 shown that the composition of the sampled aerosol is little changed by the concentration process.
31 The BYU studies listed above have shown that the concentration efficiencies for sulfate, organic
April 2002 2-50 DRAFT-DO NOT QUOTE OR CITE
-------
1 carbon (OC) and elemental carbon (EC) are comparable for a given sampling location.
2 Furthermore, the concentrations of these species and of fine particulate nitrate determined using
3 the PC-BOSS have been shown to be comparable to those determined using more conventional
4 samplers for sulfate or EC or using simpler denuder systems for OC and nitrate.
5 Ding et al. (1998a) developed a method for the determination of total n-nitroso compounds
6 in air samples and used the method to examine organic compounds formed from NOX chemistry
7 in Provo, UT (Ding et al., 1998b). In their method, n-nitroso compounds are selectively
8 decomposed to yield nitric oxide, which is then detected using chemiluminescence. From the
9 samples from Provo, they found that the majority of the n-nitroso and nitrite organic compounds
10 that were present in fine parti culate matter were semivolatile organic compounds that could be
11 evaporated from the particles during sampling. They found particulate n-nitroso compound
12 concentrations ranging between <1 and 3 nmoles/m3 and gas-phase n-nitroso compound
13 concentrations in the same range. Particulate organic nitrite concentrations were found in the
14 range of <1 to ~5 nmoles/m3, and gas-phase concentrations as high as 10 nmoles/m3 were found.
15 Turpin et al. (1993) developed a sampling system that corrects for the loss of semivolatile
16 organic compounds during sampling by removal of most of the gas phase material from the
17 particles in a diffusion separator sampling system. Unlike the previously mentioned systems,
18 wherein the particulate phase is measured directly, in the system of Turpin et al. (1993) the
19 gas-phase is measured directly. In the laminar flow system, ambient, particle-laden air enters the
20 sampler as an annular flow. Clean, particle-free air is pushed through the core inlet of the
21 separator. The clean air and ambient aerosol join downstream of the core inlet section, and flow
22 parallel to each other through the diffusion zone. Because of the much higher diffusivities for
23 gases compared to particles, the SVOC in the ambient air diffuses to the clean, core flow. The
24 aerosol exits the separator in the annular flow, and the core flow exiting the separator now
25 contains a known fraction of the ambient SVOC. Downstream of the diffusion separator, the
26 core exit flow goes into a polyurethane foam (PUF) plug, where the SVOC is collected. The
27 adsorbed gas phase on the PUF plug is extracted with supercritical fluid CO2 and analyzed by gas
28 chromatography/mass-selective detection (GC/MSD). The gas-phase SVOC is thus determined.
29 Ultimately, to determine particulate phase SVOC concentrations, the total compound
30 concentration will also be measured and the particulate phase obtained by difference. The system
31 was tested for the collection of PAH. The diffusional transport of gas-phase PAHs and particle
April 2002 2-51 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations agreed well with theory. Breakthrough was problematic for low molecular weight
2 PAHs (MW < 160). Detection limits ranged from 20 to 50 pg of injected mass for all PAHs.
3 Gundel et al. (1995) recently developed a technique for the direct determination of phase
4 distributions of semivolatile polycyclic aromatic hydrocarbons using annular denuder technology.
5 The method, called the integrated organic vapor/particle sampler (IOVPS), uses a cyclone inlet
6 with a 50% cutpoint of 2.5 //m at a sampling rate of 10 L/min. The airstream then goes through
7 two or three sandblasted glass annular denuders that are coated with ground adsorbent resin
8 material (XAD-4 was initially examined) that traps vapor-phase organics. The airstream
9 subsequently passes through a filter, followed by a backup denuder. The denuder collection
10 efficiency is high and compares well with predictions based on the diffusivity of the compounds.
11 The denuder can also be extracted to obtain gas-phase concentrations directly (Gundel and Lane,
12 1999). Particle-phase PAHs are taken to be the sum of material on the filter and XAD adsorbent
13 downstream after correction for denuder collection efficiency. The IOVPS was tested for
14 sampling semivolatile PAH in laboratory indoor air and in environmental tobacco smoke (ETS).
15 After exposure, the denuders, filters, and sorbent traps were extracted with cyclohexane (Gundel
16 et al., 1995) and analyzed for PAHs from naphthalene to chrysene using dual-fluorescence
17 detection (Mahanama et al., 1994). Recoveries from both denuders and filters were
18 approximately 70% for 30 samples. Detection limits (defined as 3 times the standard deviation
19 of the blanks) for gas-phase SVOC PAHs ranged from 0.06 ng for anthracene to 19 ng for
20 2-methylnaphthalene. The 95% confidence interval for reproduction of an internal standard
21 concentration was 6.5% of the mean value. Relative precision, from a propagation of errors
22 analysis or from the 95% confidence interval from replicate analyses of standard reference
23 material SRM 1649 (urban dust/organics), was 12% on average (8% for naphthalene to 22% for
24 fluorene). Sources of error included sampling flow rate, internal standard concentration, and
25 co-eluting peaks. Gundel and Lane (1999) reported that roughly two-thirds of particulate PAH
26 fluoranthene, pyrene, benz[a]anthracene, and chrysene were found on the postfilter denuders, so
27 that it is likely that considerable desorption from the collected particles took place.
28 Solid adsorbent-based denuder systems have been investigated by other researchers as well.
29 Bertoni et al. (1984) described the development of a charcoal-based denuder system for the
30 collection of organic vapors. Risse et al. (1996) developed a diffusion denuder system to sample
31 aromatic hydrocarbons. In their system, denuder tubes with charcoal coating and charcoal paper
April 2002 2-52 DRAFT-DO NOT QUOTE OR CITE
-------
1 precede a filter pack for particulate collection and an adsorption tube to capture particle blow-off
2 from the filter sample. Breakthrough curves for benzene, toluene, ortho-xylene, and meta-xylene
3 were developed for 60-, 90-, and 120-cm denuder tubes. The effects of relative humidity on the
4 adsorption capacities of the denuder system were examined, and it was found that the capacity of
5 the charcoal was not affected significantly by increases in relative humidity. The feasibility of
6 outdoor air sampling with the system was demonstrated.
7 Krieger and Kites (1992) designed a diffusion denuder system that uses capillary gas
8 chromatographic columns as the tubes for SVOC collection. The denuder was followed by a
9 filter to collect particles, which in turn was followed by a PUF plug to collect organic material
10 volatilizing off the filter. Denuder samples were analyzed by liquid solvent extraction (CH2C12)
11 followed by GC-mass spectrometric analysis. The PUF plugs and filters were extracted with
12 supercritical fluid extraction using supercritical N2O. Using this system, an indoor air sample
13 was found to contain primarily chlorinated biphenyls, ranging from trichlorobiphenyls (vapor
14 pressures 10'3 - 10'4 Torr at 25 °C) to octachlorobiphenyls (10'6 - 10'7 Torr). This demonstrated
15 that the sampler collects compounds with a wide range of volatility. They also found that on-line
16 desorption is successful in maintaining good chromatographic peak shape and resolution. The
17 entire method, from sample collection to the end of the chromatographic separation, took 2 h.
18 Organic acids in both the vapor and particulate phases may be important contributors to
19 ambient acidity, as well as representing an important fraction of organic particulate matter.
20 Lawrence and Koutrakis (1996a,b) used a modified Harvard/EPA annular denuder system
21 (HEADS) to sample both gas and particulate phase organic acids in Philadelphia, PA, in the
22 summer of 1992. The HEADS sampler inlet had a 2. l-//m cutpoint impactor (at 10 L/min),
23 followed by two denuder tubes, and finally a filter pack with a Teflon filter. The first denuder
24 tube was coated with potassium hydroxide (KOH) to trap gas phase organic acids. The second
25 denuder tube was coated with citric acid to remove ammonia and thus to avoid neutralizing
26 particle phase acids collected on the filter. The KOH-coated denuder tube was reported to collect
27 gas phase formic and acetic acids at better than 98.5% efficiency and with precisions of 5% or
28 better (Lawrence and Koutrakis, 1994). It was noted that for future field measurements of
29 particulate organic acids, a Na2CO3-coated filter should be deployed downstream of the Teflon
30 filter to trap organic acids that may evaporate from the Teflon filter during sampling.
31
April 2002 2-53 DRAFT-DO NOT QUOTE OR CITE
-------
1 Role of the Collection Media
2 The role of the collection media was recently examined in a study conducted in Seattle
3 (Lewtas et al., 2001). In that study, the influence of denuder sampling methods and filter
4 collection media on the measurement of SVOC associated with PM2 5 was evaluated. Activated
5 carbon and XAD collection media were used in diffusion denuders and impregnated back-up
6 filters in two different samplers, the Versatile Air Pollution Sampler (VAPS) and the PC-BOSS.
7 XAD-coated glass annular denuders and charcoal-impregnated cellulose fiber (GIF) filter
8 denuders also were used. GIF filters also were compared to XAD-coated quartz filters as backup
9 filter collection media. Lewtas et al. (2001) found that the two denuder types resulted in an
10 equivalent measurement of particulate organic carbon and particle mass. The carbon-coated
11 denuders in the BOSS sampler were more efficient than the XAD-coated denuders for the
12 collection of the more highly volatile organic compounds (MHVOC). Lewtas et al. (2001)
13 concluded that this MHVOC that is collected in the carbon-coated BOSS denuder does not
14 contribute substantially to the particle mass or to the SVOC measured as OC on quartz filters.
15 However, this MHVOC would be captured in the carbon impregnated filters placed behind the
16 quartz filters, so that, in the XAD denuder configuration, the captured MHVOC would cause a
17 higher OC concentration and an overestimation of the SVOC.
18 Some of the recent research in denuder technology also has focused on reduction in the size
19 of the denuder, optimization of the residence time in the denuder, understanding the effect of
20 diffusion denuders on the positive quartz filter artifact, identifying changes in chemical
21 composition that occur during sampling, determining the effects of changes in temperature and
22 relative humidity, and identifying possible losses by absorption in coatings.
23
24 Reducing the Size of Denuders
25 The typical denuder configuration is an annular diffusion denuder tube of significant length
26 (e.g., 26.5 cm for 10 L/min [Koutrakis et al., 1988]). A more compact design based on a
27 honeycomb configuration was shown to significantly increase the capacity (Koutrakis et al.,
28 1993). However, in intercomparisons with an annular denuder/filter pack system (Koutrakis
29 et al., 1988), significant losses of ammonia and nitric acid were observed for the honeycomb
30 configuration and were attributed to the large inlet surface area and long sample residence time
31 of the honeycomb design relative to the annular denuder system. Sioutas et al. (1996a)
April 2002 2-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 subsequently designed a modified glass honeycomb denuder/filter pack sampler (HDS) with an
2 inlet that minimizes vapor losses on the inlet surfaces. The modified HDS has reduced inlet
3 surfaces and decreased residence time for sampled gases (NH3 and HNO3) compared to its
4 predecessor (Sioutas et al., 1994a). Sioutas et al. (1996b) also tested various inlet materials
5 (glass, PFA, and polytetrafluoroethylene [PTFE]) in laboratory tests and found that a PTFE
6 Teflon coated inlet minimized loss of sampled gases (1 to 8% loss of HNO3 observed, and -4 to
7 2% loss of NH3 observed). The highest inlet losses were observed for FINO3 lost to PFA surfaces
8 (14 to 25%). The modified HDS was tested in laboratory and field tests and found to agree
9 within 10% with the annular denuder system.
10
11 Residence Time in the Denuder
12 The efficiency of a diffusion denuder sampler for the removal of gas phase material can be
13 improved by increasing the residence time of the sampled aerosol in the denuder. However, the
14 residence time can only be increased within certain limits. Because the diffusion denuder
15 reduces the concentration of gas-phase semivolatile organic material, semivolatile organic matter
16 present in the particles passing through the denuder will be in a thermodynamically unstable
17 environment and will tend to outgas SVOC during passage through the denuder. The residence
18 time of the aerosol in the denuder, therefore, should be short enough to prevent significant loss of
19 particulate phase SVOC to the denuder. Various studies have suggested that the residence time
20 in the denuder should be less than about 2 s (Gundel and Lane, 1999; Kamens and Coe, 1997;
21 Kamens et al., 1995). The residence times in the various denuder designs described by Gundel
22 and Lane (1999) are from 1.5 to 0.2 s. The equilibria and evaporation rates are not as well
23 understood for organic components as they are for NH4NO3 (Zhang and McMurry, 1987, 1992;
24 Hering and Cass, 1999).
25
26 Effect of Diffusion Denuders on the Positive Quartz Filter Artifact
27 The adsorption of organic compounds by a second quartz filter has been shown to be
28 reduced, but not eliminated, in samples collected in the Los Angeles Basin if a multi-channel
29 diffusion denuder with quartz filter material as the denuder collection surface preceded the quartz
30 filters (Fitz, 1990). This artifact can be further reduced by the use of activated charcoal as the
31 denuder surface and the use of a particle concentrator to reduce the amount of gas phase organic
April 2002 2-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 compounds relative to condensed phase organic compounds (Cui et al., 1997, 1998; Eatough,
2 1999). Recent experiments (Cotham and Bidleman, 1992; Cui et al., 1998; Eatough et al., 1995,
3 1996) have shown that the quartz filter artifact can result both from the collection of gas phase
4 organic compounds and from the collection of semivolatile organic compounds lost from
5 particles during sampling. Thus, results available to date suggest that both a "positive" and a
6 "negative" artifact can be present in the determination of particulate phase organic compounds
7 using two tandem quartz filters.
8 The importance of the adsorption of organic vapors on filters or PM relative to the
9 volatilization of organic compounds from PM collected on a filter continues to be a topic of
10 active debate. The relative importance of positive and negative artifacts will be different for
11 denuded and undenuded filters; will depend on face velocity, sample loading, and the vapor
12 pressures of the compounds of interest; and may vary with season and location because of
13 variations in the composition of volatile and semivolatile organic material. Evidence exists for
14 substantial positive and negative artifacts in the collection of organic PM.
15
16 Changes in Chemical Composition During Sampling
17 The use of sampling systems designed to correctly identify the atmospheric gas and
18 parti culate phase distributions of collected organic material has been outlined above.
19 An additional sampling artifact that has received little consideration in the collection of
20 atmospheric samples is the potential alteration of organic compounds as a result of the sampling
21 process. These alterations appear to result from the movement of ambient air containing
22 oxidants and other reactive compounds past the collected particles. The addition of NO2
23 (
-------
1 Similar oxidation of aldehydes and peroxyacetylnitrate (PAN) during sampling has been
2 observed (Grosjean and Parmar, 1990). Collected PAH compounds can be oxygenated or
3 nitrated on a filter (Davis et al., 1987; Lindskog and Brorstrom-Lunden, 1987), but 1-nitropyrene
4 has been shown to be resistant to additional nitration (Grosjean, 1983). These various chemical
5 transformations of collected organic compounds can be eliminated by removal of the gas phase
6 oxidants, NOX, HNO3, etc., or by reaction or absorption prior to collection of the particles (Ding,
7 1998a,b; Grosjean and Parmar, 1990; Parmar and Grosjean, 1990; Pellizzari and Krost, 1984;
8 Williams and Grosjean, 1990). The BOSS denuder should be effective in eliminating most of the
9 chemical transformation artifacts because reactive gases are removed by the charcoal denuder
10 that precedes the particle collection filter.
11
12 Temperature and Relative Humidity Effects
13 The problems of sampling artifacts associated with SVOC adsorption and evaporation are
14 compounded by temperature and relative humidity effects (Pankow and Bidleman, 1991; Pankow
15 et al., 1993; Falconer et al., 1995; Goss and Eisenreich, 1997). Effects of temperature on the
16 partitioning of PAH were examined by Yamasaki et al. (1982), who found that the partition
17 coefficient (PAHvapor/PAHpart) was inversely related to temperature and could be described using
18 the Langmuir adsorption concept. The dissociation of ammonium nitrate aerosol is also a
19 function of temperature. Bunz et al. (1996) examined the dissociation and subsequent
20 redistribution of NH4NO3 within a bimodal distribution using a nine-stage low-pressure Berner
21 impactor followed by analysis by ion chromatography and found a strong temperature
22 dependency on the redistribution. Bunz et al. (1996) found that at lower temperatures (below
23 10 °C) there was little change in the aerosol size distribution. At temperatures between 25 and
24 45 °C, however, the lifetime of NH4NO3 particles decreases by more than a factor of 10, and size
25 redistribution, as measured by average ending particle diameter, increased more for higher
26 temperatures than for lower temperatures.
27 The effects of relative humidity on the sorption of SVOC on particles are not well
28 understood. In a series of laboratory experiments, Goss and Eisenreich (1997) examined the
29 sorption of both nonpolar (hydrocarbons and chlorinated hydrocarbons) and polar (ethyl ether
30 and acetone) volatile organic compounds onto combustion soot particles as a function of
31 temperature and relative humidity. The soot particles used in their experiments were collected
April 2002 2-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 from oil furnaces and contained 60% (w/w) iron sulfate (water-soluble fraction) and 9% (w/w)
2 elemental and organic carbon. They found that, for all compounds, the sorption of VOC onto
3 soot particles decreased with increasing relative humidity over the range of 10 to 95%. They also
4 observed hysteresis in the relative humidity dependency, with sorption coefficients at a given
5 relative humidity higher when the RH is being increased than when the RH is being decreased.
6 The sorption coefficients were fit with an exponential function to the RH so that the slope of the
7 regression line would provide a measure of the influence of relative humidity. Based on the
8 magnitude of the slope, they concluded that the RH-dependency of sorption was stronger for
9 water-soluble organic compounds.
10 In another study by Jang and Kamens (1998), humidity effects on gas-particle partitioning
11 of S VOC were examined using outdoor environmental chambers and the experimentally
12 determined partitioning coefficients were compared to theoretical values. They examined the
13 partitioning of SVOC onto wood soot, diesel soot, and secondary aerosols and concluded that
14 "the humidity effect on partitioning was most significant for hydrophobic compounds adsorbing
15 onto polar aerosols." Although these two studies seem to be contradictory, on closer
16 examination, it is difficult to compare the two studies for several reasons. The experiments
17 conducted by Jang and Kamens (1998) were conducted in outdoor chambers at ambient
18 temperatures and humidities. Their model was for absorptive partitioning of SVOC on
19 liquid-like atmospheric particulate matter. In contrast, the results of Goss and Eisenreich (1997)
20 were obtained from a gas chromatographic system operated at 70 °C higher than ambient
21 conditions. The model of Goss and Eisenreich (1997) was for adsorptive partitioning of VOC on
22 solid-like atmospheric parti culate matter. In the study of Jang and Kamens (1998), calculated
23 theoretical values for water activity coefficients for diesel soot were based on an inorganic salt
24 content of 1 to 2%; whereas, the combustion particles studied by Goss and Eisenreich (1997)
25 contained 60% water-soluble, inorganic salt content. Jang and Kamens (1998) obtained their
26 diesel soot from their outdoor chamber, extracted it with organic solvent (mixtures of hexane and
27 methylene chloride), and measured the organic fraction. The resulting salt content of 2% of the
28 particulate matter studied in Jang and Kamens (1998) is enough to affect water uptake but
29 presumably not to affect the sorption partitioning of organics.
30
31
April 2002 2-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 Impactor Coatings
2 Impactors are used as a means to achieve a size outpoint and as particle collection surfaces.
3 Particles collected on impactors are exposed to smaller pressure drops than filter-collected
4 particles, making them less susceptible to volatile losses (Zhang and McMurry, 1987). However,
5 size resolution can be affected by bounce when samples are collected at low humidities (Stein
6 et al., 1994). There are other sources of error inherent in some of the currently acceptable
7 practices that could potentially affect paniculate mass concentration measurements and that will
8 surely become even more important as more emphasis is placed on chemical speciation. Allen
9 et al. (1999a) reported that the practice of greasing impaction substrates may introduce an artifact
10 from the absorption of semivolatile species from the gas phase by the grease because the grease
11 could artificially increase the amount of PAHs and other organic compounds attributed to the
12 aerosol. Allen et al. (1999a) offer several criteria to ensure that this absorption artifact is
13 negligible, including selecting impaction oils in which analytes of interest are negligibly soluble
14 and ensuring that species do not have time to equilibrate between the vapor and oil phases
15 (criterion is met for nonvolatile species). They recommend using oiled impaction substrates only
16 if the absorption artifact is negligible as determined from these criteria. Application of greases
17 and impaction oils for preventing or reducing bounce when sampling with impactors is not
18 suitable for carbon analysis because the greases contain carbon (Vasilou et al., 1999).
19 Kavouras and Koutrakis (2001) investigated the use of polyurethane foam (PUF) as a
20 substrate for conventional inertial impactors. The PUF impactor substrate is not rigid like the
21 traditional impactor substrate so particle bounce and reentrainment artifacts are reduced
22 significantly. Kavouras and Koutrakis (2001) found that the PUF impaction substrate resulted in
23 a much smaller 50% cut point at the same flow rate and Reynolds number. Moreover, the lower
24 50% cut point was obtained at a lower pressure drop than with the conventional substrate, which
25 could lead to a reduction of artifact vaporization of semivolatile components.
26
27 2.2.3.3 Particle-Bound Water
28 It is generally desirable to collect and measure ammonium nitrate and semivolatile organic
29 compounds. However, for many measurements of suspended particle mass, it is desirable to
30 remove the particle-bound water before determining the mass. In other situations it may be
31 important to know how much of the suspended particle's mass or volume results from particle-
April 2002 2-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 bound water. The water content of PM is significant and highly variable. Moreover, there is
2 significant hysteresis in the water adsorption-desorption pathways (Seinfeld and Pandis, 1998),
3 further complicating the mass measurement. Figures 2-8 and 2-9 show the change in diameter of
4 sulfate particles as a function of relative humidity. Figure 2-8 also shows the difference between
5 deliquescence and crystallization points.
6 Pilinis et al. (1989) calculated the water content of atmospheric particulate matter above
7 and below the deliquescent point. They predicted that aerosol water content is strongly
8 dependent on composition and concluded from their calculations that liquid water could
9 represent a significant mass fraction of aerosol concentration at relative humidities above 60%.
10 Since then, a few researchers have attempted to measure the water content of atmospheric
11 aerosol. Most techniques have focused on tracking the particle mass as the relative humidity is
12 changed and are still in the development phase. There have been only a few demonstrations
13 using actual ambient aerosol to date. Of interest, in particular, is the development of the Tandem
14 Differential Mobility Analyzer (TDMA) and its applications in investigations of the effects of
15 relative humidity on particle growth.
16 Lee et al. (1997) examined the influence of relative humidity on the size of atmospheric
17 aerosol using a TDMA coupled with a scanning mobility particle sizer (SMPS). They reported
18 that the use of the TDMA/SMPS system allowed for the abrupt size changes of aerosols at the
19 deliquescence point to be observed precisely. They also reported that at relative humidities
20 between 81 and 89% the water content of ammonium sulfate aerosols (by mass) ranged from
21 47 to 66%.
22 Andrews and Larson (1993) investigated the interactions of single aerosol particles coated
23 with an organic film within a humid environment. Using an electrodynamic balance, they
24 conducted laboratory experiments in which sodium chloride and carbon black particles were
25 coated with individual organic surfactants (intended to simulate the surface-active, organic films
26 that many atmospheric aerosol particles may exhibit) and their water sorption curves were
27 examined. Their results showed that when ordinarily hydrophobic carbon black particles were
28 coated with an organic surfactant, they sorbed significant amounts of water (20 to 40% of the dry
29 mass of the particle).
30 Liang and Chan (1997) developed a fast technique using the electrodynamic balance to
31 measure the water activity of atmospheric aerosols. In their technique, the mass of a levitated
April 2002 2-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 particle is determined as the particle either evaporates or grows in response to a step change in
2 the relative humidity. Their technique was demonstrated using laboratory experiments with
3 NaCl, (NH4)2SO4, NaNO3, and (NH4)2SO4/NH4NO3 solutions. They concluded that one of the
4 advantages of their fast method is the ability to measure the water activity of aerosols containing
5 volatile solutes such as ammonium chloride and some organics.
6 Mclnnes et al. (1996) measured aerosol mass concentration, ionic composition, and
7 associated water mass of marine aerosol over the remote Pacific Ocean. The mass of
8 particle-bound water was determined by taking the difference between the mass obtained at 48%
9 RH and at 19% RH, assuming the aerosol particles were dry at 19% RH. Based on a comparison
10 of the remote Pacific aerosol to aerosol collected at a site at the marine/continental interface of
11 the Washington coast, the amount of water associated with the aerosol was observed to be a
12 function of the ammonium to sulfate ratio. They found that the amount of water associated with
13 the submicrometer aerosol comprised 29% of the total aerosol mass collected at 47% RH and
14 9% of the total mass at 3 5% RH.
15 Ohta et al. (1998) characterized the chemical composition of atmospheric fine particles
16 (50% cut point of 2 //m) in Sapporo, Japan, and as part of their measurements, determined the
17 water content using the Karl Fischer method (Meyer and Boyd, 1959). After exposing a Teflon
18 filter, a portion of the filter was equilibrated at 30% RH for 24 h. Then the filter piece was
19 placed in a water evaporator heated at 150 °C, vaporizing the particle-bound water. The vapor
20 evolved was analyzed for water in an aqua-counter where it was titrated coulometrically in Karl
21 Fischer reagent solution (containing iodine, sulfur, and methanol). The accuracy of the aqua-
22 counter is ±1 mg. Using this technique, they determined that the water content of the particles
23 ranged from 0.4 to 3.2% of the total particulate mass (at RH < 30%). This represents a smaller
24 portion of water compared to their previous reported values (Ohta and Okita, 1990) that were
25 determined by calculation at RH of 50%.
26 Speer et al. (1997) developed an aerosol liquid water content analyzer (LWCA) in which
27 aerosol samples are collected on PTFE filters and then placed in a closed chamber in which the
28 relative humidity is closely controlled. The aerosol mass is monitored using a beta-gauge, first as
29 the relative humidity is increased from low RH to high RH, and then as the RH is decreased
30 again. They demonstrated the LWCA on laboratory-generated aerosol and on an ambient PM25
31 sample collected in Research Triangle Park, NC. The ambient aerosol sample was also analyzed
April 2002 2-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 for chemical constituents. It is interesting to note that, although their laboratory-generated
2 (NH4)2SO4 aerosol demonstrated a sharp deliquescent point, their atmospheric aerosol, which
3 was essentially (NH4)2SO4, did not show a sharp deliquescent point.
4 Hygroscopic properties of aerosols have been studied from the viewpoint of their ability to
5 act as condensation nuclei. The hygroscopic properties of fresh and aged carbon and diesel soot
6 particles were examined by Weingartner et al. (1997) who found that fresh, submicron-size
7 particles tended to shrink with increasing relative humidity because of a structural change.
8 Lamm el and Novakov (1995) found, in laboratory studies, that the hygroscopicity of soot
9 particles could be increased by chemical modification and that the cloud condensation nucleation
10 characteristics of diesel soot were similar to those of wood smoke aerosol.
11 The results of several of the above studies in which aerosol water content as a function of
12 relative humidity was determined are summarized in Figure 2-13. In this figure, the results of
13 Lee et al. (1997), Mclnnes et al. (1996), and Ohta et al. (1998) are included. Relative humidity
14 ranged from 9%, at which the aerosol water content was assumed to be zero (Mclnnes et al.,
15 1996), to 89%, at which the aerosol water content was determined to be 66% by mass (Lee et al.,
16 1997). Koutrakis et al. (1989) and Koutrakis and Kelly (1993) also have reported field
17 measurements of the equilibrium size of atmospheric sulfate particles as a function of relative
18 humidity and acidity.
19 The effects of relative humidity on particle growth were also examined in several studies.
20 Fang et al. (1991) investigated the effects of flow-induced relative humidity changes on particle
21 cut sizes for aqueous sulfuric acid particles in a multi-nozzle micro-orifice uniform deposit
22 impactor (MOUDI). Laboratory experiments were conducted in which polydisperse sulfuric acid
23 aerosols were generated and the RH was adjusted. The aerosols were analyzed by a differential
24 mobility analyzer. Fang et al. (1991) observed that for inlet RH less than 80%, the cut sizes for
25 the sulfuric acid aerosols were within 5% of that for nonhygroscopic particles except at the stage
26 for which the cut size was 0.047 //m where the cut size was 10.7% larger than the
27 nonhygroscopic particle cut size. They concluded that flow-induced RH changes would have
28 only a modest effect on MOUDI cut sizes at RH < 80%.
29 Hitzenberger et al. (1997) collected atmospheric aerosol in the size range of 0.06 to 15 //m
30 in Vienna, Austria, using a nine-stage cascade impactor and measured the humidity-dependent
31 water uptake when the individual impaction foils were exposed to high RH. They observed
April 2002 2-62 DRAFT-DO NOT QUOTE OR CITE
-------
CD
s
T3
C
O
CO
V
o
CO
Q_
0
100
90-
80
70-
60-
50-
g 40-
co 30-
•g 2o^
£ 10-1
• Mclnnes etal., 1996
• Lee etal., 1997
A Ohtaetal., 1998
10 20 30 40 50 60
Relative Humidity, %
70
80
90
100
Figure 2-13. Aerosol water content expressed as a mass percentage, as a function of
relative humidity.
1 particle growth with varying growth patterns. Calculated extinction coefficients and single
2 scattering albedo increased with humidity.
3 Hygroscopic properties, along with mixing characteristics, of submicrometer particles
4 sampled in Los Angeles, CA, during the summer of 1987 SCAQS study and at the Grand
5 Canyon, AZ, during the 1990 Navajo Generating Station Visibility Study were reported by Zhang
6 et al. (1993). They used a tandem differential mobility analyzer (TDMA; McMurry and
7 Stolzenburg, 1989) to measure the hygroscopic properties for particles in the 0.05- to 0.5-//m
8 range. In their experimental technique, monodisperse particles of a known size are selected from
9 the atmospheric aerosol with the first DMA. Then, the relative humidity of the monodisperse
10 aerosol is adjusted, and the new particle size distribution is measured with the second DMA.
11 At both sites, they observed that monodisperse particles could be classified according to "more"
12 hygroscopic and "less" hygroscopic. Aerosol behavior observed at the two sites differed
13 markedly. Within the experimental uncertainty (±2%) the "less" hygroscopic particles sampled
14 in Los Angeles did not grow when the RH was increased to 90%; whereas at the Grand Canyon,
April 2002
2-63
DRAFT-DO NOT QUOTE OR CITE
-------
1 the growth of the "less" hygroscopic particles varied from day to day, but ranged from near 0 to
2 40% when the RH was increased to 90%. The growth of the "more" hygroscopic particles in
3 Los Angeles was dependent on particle size (15% at 0.05 //m to 60% at 0.5 //m); whereas at the
4 Grand Canyon, the "more" hygroscopic particles grew by about 50% with the growth not varying
5 significantly with particle size. By comparison of the TDMA data to impactor data, Zhang et al.
6 (1993) surmised that the more hygroscopic particles contained more sulfates and nitrates while
7 the less hygroscopic particles contained more carbon and crustal components.
8 Although most of the work to date on the hygroscopic properties of atmospheric aerosols
9 has focused on the inorganic fraction, the determination of the contribution of particle-bound
10 water to atmospheric particulate mass is greatly complicated by the presence of organics. The
11 effect of RH on adsorption of semivolatile organic compounds is discussed elsewhere in this
12 chapter. Saxena et al. (1995) observed that particulate organic compounds also can affect the
13 hygroscopic behavior of atmospheric particles. They idealized the organic component of aerosol
14 as containing a hydrophobic fraction (high-molecular weight alkanes, alkanoic acids, alkenoic
15 acids, aldehydes, and ketones) and a hydrophilic fraction (e.g., lower molecular weight
16 carboxylic acids, dicarboxylic acids, alcohols, aldehydes, etc.) that would be likely to absorb
17 water. They then analyzed data from a tandem differential mobility analyzer in conjunction with
18 particle composition observations from an urban site (Claremont, CA) and from a nonurban site
19 (Grand Canyon) to test the hypothesis that, by adding particulate organics to an inorganic aerosol,
20 the amount of water absorbed would be affected, and the effect could be positive or negative,
21 depending on the nature of the organics added. They further presumed that the particulate
22 organic matter in nonurban areas would be predominantly secondary and thus hydrophilic,
23 compared to the urban aerosol that was presumed to be derived from primary emissions and thus
24 hydrophobic in nature. Their observations were consistent with their hypothesis, in that at the
25 Grand Canyon, the presence of organics tended to increase the water uptake by aerosols; whereas
26 at the Los Angeles site, the presence of organics tended to decrease water uptake.
27 Peng and Chan (2001) also recently studied the hygroscopic properties of nine water
28 soluble organic salts of atmospheric interest using an electrodynamic balance operated at 25 °C.
29 Salts studied included sodium formate, sodium acetate, sodium succinate, sodium pyruvate,
30 sodium methanesulfonate, sodium oxalate, ammonium oxalate, sodium malonate, and sodium
31 maleate. They observed that hygroscopic organic salts have a growth factor of 1.76-2.18 from
April 2002 2-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 RH= 10-90%, comparable to that of typical hygroscopic inorganic salts such as NaCl and
2 (NH4)2S04.
3 Nonequilibrium issues may be important for the TDMA, as well as for other methods of
4 measuring water content. Although approach to equilibrium when the RH is increased is
5 expected to be rapid for pure salts, it may be much slower for aerosols containing a complex mix
6 of components (Saxena et al., 1995). For example, if an aerosol contains an organic film or
7 coating, that film may impede the transport of water across the particle surface, thus increasing
8 the time required for equilibrium (Saxena et al., 1995). Insufficient time to achieve equilibrium
9 in the TDMA could result in underestimation of the water content.
10
11 2.2.4 U. S. Environmental Protection Agency Monitoring Methods
12 2.2.4.1 The Federal Reference Methods for Measurement of Equilibrated Mass for
13 PM10,PM25,andPM1025
14 In 1997, EPA promulgated new standards for PM2 5 to address fine-fraction thoracic
15 particles and retained with minor revisions the 1987 PM10 standards to continue to address
16 coarse-fraction thoracic particles (Federal Register, 1997). In partial response to numerous
17 challenges to these standards, the U.S. Court of Appeals for the District of Columbia Circuit in
18 American Trucking Association v. EPA, 175 F. 3d 1027 (U.S. Court of Appeals, D.C. Cir. 1999)
19 found "ample support" for regulating coarse-fraction particles but revoked the revised PM10
20 standards (leaving in effect the 1987 PM10 standards) on the basis of PM10 being a "poorly
21 matched indicator for coarse particulate pollution" because PM10 includes fine particles.
22 Consistent with this specific aspect of the Court's ruling, which EPA did not appeal, EPA is now
23 considering use of PM10_2 5 as the indicator for coarse-fraction thoracic particles, in conjunction
24 with PM2 5 standards that address fine-fraction thoracic particles. Thus, EPA is now developing a
25 Federal Reference Method for the measurement of PM10_2 5.
26
27 2.2.4.1.1 PM10
28 The FRM specified for measuring PM10 (Code of Federal Regulations, 2001a,b) has been
29 discussed in previous PM AQCD's and will only be mentioned briefly. The PM10 FRM defines
30 performance specifications for samplers in which particles are inertially separated with a
31 penetration efficiency of 50% at an aerodynamic diameter (Da) of 10 ± 0.5 //m. The collection
April 2002 2-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 efficiency increases to ~ 100% for smaller particles and drops to ~ 0% for larger particles.
2 Particles are collected on filters and mass concentrations are determined gravimetrically.
3 Instrument manufacturers are required to demonstrate through field tests a measurement
4 precision for 24-h samples of ± 5 //g/m3 for PM10 concentrations below 80 //g/m3 and 7% above
5 this value. A number of samplers have been designated as PM10 reference samplers. The TEOM
6 and several beta gauge samplers with 1-h time resolution have been designated as automated
7 equivalent methods (U.S. Environmental Protection Agency, 2001).
8
9 2.2.4.1.2 PM25
10 As opposed to the performance-based FRM standard for PM10, the FRM for PM2 5 (Code of
11 Federal Regulations, 200la) specifies certain details of the sampler design, as well as of sample
12 handling and analysis, whereas other aspects have performance specifications (Noble et al.,
13 2001). The PM2 5 FRM sampler consists of a PM10 inlet/impactor, a PM2 5 impactor with an oil-
14 soaked impaction substrate to remove particles larger than 2.5 //m Da, and a 47-mm PTFE filter
15 with a particle collection efficiency greater than 99.7%. The sample duration is 24 h, during
16 which time the sample temperature is not to exceed ambient temperatures by more than 5 °C.
17 A schematic diagram of the PM25 FRM sample collection system is shown in Figure 2-14. After
18 collection, samples are equilibrated for 24 h at temperatures in the range of 20 to 23 °C (± 2 °C)
19 and at relative humidities in the range of 30 to 40% (± 5%). The equilibration tends to reduce
20 particle-bound water and stabilizes the filter plus sample weight. Filters are weighed before and
21 after sampling under the same temperature and relative humidity conditions. For sampling
22 conducted at ambient relative humidity less than 30%, mass measurements at relative humidities
23 down to 20% are permissible (Code of Federal Regulations, 200la).
24 The PM10 inlet specified for the PM2 5 FRM is modified from a previous low flow-rate PM10
25 inlet that was acceptable in both EPA-designated reference and equivalent PM10 methods. The
26 modification corrects a flaw that was reported for the previous sampler, in that under some
27 meteorological conditions, the inlet may allow precipitation to penetrate the inlet. The
28 modification includes a larger drain hole, a one-piece top plate, and louvers. Tolocka et al.
29 (2001a) evaluated the performance of this modified inlet in a series of wind tunnel experiments.
30 The modified inlet was found to provide a size out comparable to the original inlet, for both
31 PM2 5 and PM10 sampling. Since the modification did not change the characteristics of the size
April 2002 2-66 DRAFT-DO NOT QUOTE OR CITE
-------
Ambient
aerosol
sampling
inlet
fractionator
Downtube
WINS
PM25
fractionator
Figure 2-14. Schematic diagram of the sample collection portion of the PM2 5 FRM
sampler.
Source: Noble et al. (2001).
1 cut, the modified inlet may be substituted for the original inlet as part of a reference or equivalent
2 method for PM10 and PM2 5 (Tolocka et al., 200la).
3 WINS Impactor. Design and calibration of the EPA PM2 5 Well Impactor Ninety-Six
4 (WINS) is given by Peters et al. (200la). The WINS impactor was designed to be deployed
5 downstream of the Graseby-Andersen 246B PM10 inlet as part of a sampler operating at a flow
6 rate of 16.7 L/m. The WINS is pictured in Figure 2-15. The PM25 inlet consists of a single jet,
7 round hole, with the jet exit directed toward an impaction surface that is comprised of a 37 mm
8 diameter glass fiber filter immersed in 1 mL of low volatility diffusion pump oil (i.e., the well).
9 Particles not having enough inertia to be removed by the impactor are captured downstream on
10 the sample collection filter. This design was selected to minimize impactor overloading that
April 2002
2-67
DRAFT-DO NOT QUOTE OR CITE
-------
PM-10 Aerosol
from Inlet
Nozzle
Collection cup
with antispill ring
Impaction surface:
filter immersed in 1 mL
Dow Corning 704
diffusion pump oil
J L
I
I
I
PM-2
samp
filter
Figure 2-15. Schematic view of the final design of the WINS.
Source: Peters et al. (200la).
1 would otherwise result in particle bounce. The oil wicks through the particulate deposit on the
2 impactor to provide a continuously wetted surface for impaction. The penetration curve
3 indicated a 50% cutpoint of 2.48 //m Da with a geometric standard deviation of 1.18%
4 (Figure 2-16).
5 The WINS separator was evaluated for its loading characteristics (Vanderpool et al., 2001)
6 by monitoring the performance after repeated operation in an artificially generated, high
7 concentration, coarse-mode aerosol composed of Arizona Test Dust, as well as in the field in
8 Rubidoux, Phoenix, Philadelphia, Research Triangle Park, and Atlanta. In the wind tunnel
9 experiments, the WINS performance was found to be a monotonic function of loading. A minus
10 5% bias in the PM2 5 measurement resulted from a coarse particulate loading of approximately
11 16mg. This negative bias was due to a slight reduction in the separator cutpoint. It was also
April 2002
2-68
DRAFT-DO NOT QUOTE OR CITE
-------
.0
"ro
a)
c
CD
CL
100
80
60
40
20
-
-
-
=
-
1
Ts
0 2
WINS Ft
O f
* F
\
0 3
nal Version
VINS Best Fit Line
erosizer Detector
luorometer Detector
V ,
I I rl I I I rl
0 4.
Aerodynamic Diameter (pm)
Figure 2-16. Evaluation of the final version of the WINS.
Source: Peters etal. (200la).
1 found that the predictable results from the controlled laboratory experiments could not be
2 extrapolated to field settings and that the WINS performance was more sensitive to the impactor
3 loading in the field than it was in experiments with the single component aerosol. Significant
4 particle bounce was not observed in either the laboratory or the field experiments. Vanderpool
5 et al. (2001) conclude that their study supports the recommendation that the FRM WINS wells
6 should be replaced after every 5 days of 24-h operation (U.S. Environmental Protection Agency,
7 1998).
8 A detailed sensitivity study of the WINS impactor was conducted (Vanderpool et al., 2001)
9 in which the effects on the impactor performance of a number of parameters were examined.
10 The results of this study are summarized in Table 2-3.
April 2002
2-69
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2-3. SUMMARY OF SENSITIVITY STUDIES OF WINS IMPACTOR
PERFORMANCE
Parameter
Amount of variance Cutpoint variation
PM2 5 mass concentration bias
Manufacturing tolerances
on WINS components
Flow control biases
Specified tolerances 0.05 micrometers
4%
0.05 micrometers
Cutpoint shift partially offset
volume bias
T and P measurement
Diffusion oil volume
Impactor loading
Ambient P variations
Air Properties
Impactor oil crystallization
Impactor oil viscosity
Allowable ambient
0.75 ml to 3 ml
After 5 24 h events
OC
-20 C
-35 C
+ 0.02 micrometers
No effect
-0.07 micrometers
Negligible
2.40 micrometers
No effect
No effect
Need to change WINS
+ 0.4%
<1.5%
Negligible
NA
No effect
No effect
more frequently than every 5 days
Source: Vanderpool et al. (2001).
1 The regulations also allow for Class I, n, and in equivalent methods for PM2 5 (Code of
2 Federal Regulations, 200Ic). Class I equivalent methods use samplers with relatively small
3 deviations from the sampler described in the FRM. Class II equivalent methods include "all
4 other PM25 methods that are based upon 24-h integrated filter samplers that are subjected to
5 subsequent moisture equilibration and gravimetric mass analysis." Class in equivalent methods
6 include non-filter-based methods such as beta attenuation, harmonic oscillating elements, or
7 nephelometry (McMurry, 2000). As of July 2001, 11 PM25 samplers (listed in Table 2-4) had
8 been tested and led to the conclusion that the PM10 sampling systems can be designed such that
9 concentration measurements are precise to ±10%. For PM25, cutpoint tolerances are not
10 expected to affect the mass concentration as much as for PM10, because the 2.5 //m Da cutpoint
11 generally occurs near a minimum in the mass distribution (e.g., Figure 2-5).
12 The PM25 mass concentration will be affected, on the other hand, by other sampling issues
13 mentioned but not discussed extensively in the previous 1996 PM AQCD (U.S. Environmental
14 Protection Agency, 1996a). These issues have been discussed earlier in this chapter and include
15 gas/particle, particle/particle, and particle/substrate interactions for sulfates and nitrates (e.g.,
April 2002
2-70
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 2-4. PM2 5 SAMPLERS CURRENTLY DESIGNATED AS FRMs FOR
PM?, MASS CONCENTRATIONS
Sampler
Manufacturer
Description
Federal Register Reference
RAAS2.5-100
RAAS2.5-300
RAAS2.5-200
Partisol 2000
Partisol-Plus 2025
Partisol 2000 audit
PQ200
PQ 200A
605 CAPS
MASS 100
MASS 300
Andersen Instruments
Andersen Instruments
Andersen Instruments
Rupprecht & Patashnick
Rupprecht & Patashnick
Rupprecht & Patashnick
BGI, Inc.
BGI, Inc.
ThermoEnvironmental Instruments
URC
URC
FRM single
FRM sequential
FRM audit
FRM single
FRM sequential
FRM audit
FRM single
FRM audit
FRM single
FRM single
FRM sequential
Vol. 63, p. 31991,6/11/98
Vol. 63, p. 31991,6/11/98
Vol. 64, p. 12167,3/11/99
Vol. 63, p. 18911,4/16/98
Vol. 63, p. 18911,4/16/98
Vol. 64, p. 19153,4/19/99
Vol. 63, p. 18911,4/16/98
Vol. 63, p. 18911,4/16/98
Vol. 63, p. 58036, 10/29/98
Vol. 65, p. 26603, 05/08/00
Vol. 65, p. 26603, 05/08/00
Source: Peters et al. (200Ib); U.S. Environmental Protection Agency (2001).
1 Appel et al., 1984), volatilization losses of nitrates (Zhang and McMurry, 1992), semivolatile
2 organic compound artifacts (e.g., Eatough et al., 1993), and relative humidity effects (e.g., Keeler
3 etal., 1988).
4 Several studies now have been reported in which the FRM was collocated with other PM2 5
5 samplers in intercomparison studies. During the Aerosol Research and Inhalation Epidemiology
6 Study (ARIES), several PM2 5 samplers were collocated at a mixed industrial-residential site near
7 Atlanta, GA (Van Loy et al., 2000). These samplers included a standard PM25 FRM, a TEOM
8 with Nafion drier, a particulate composition monitor (PCM; Atmospheric Research and Analysis,
9 Gary, NC), a medium-volume (113 L/min flow rate) fine particle (PM25) and semivolatile
10 organic compound (i.e., a filter followed by a solid adsorbent) sampler, operated by the Desert
11 Research Institute, a HEADS sampler, and a dichotomous sampler for coarse PM. The PCM
12 sampler has three channels, all of which have PM10 cyclone inlets. The first two channels both
13 have two denuders preceding a 2.5-//m WINS impact and filter packs. The first denuder is
14 coated with sodium carbonate to remove acid gases, and the second is coated with citric acid to
April 2002
2-71
DRAFT-DO NOT QUOTE OR CITE
-------
1 remove ammonia. The third channel has a carbon coated parallel-plate denuder preceding the
2 WINS impactor. Measurements of 24-h mass from the FRM, PCM, and TEOM samplers, as
3 well as reconstructed PM2 5 mass (RPM), were compared for a 12-mo period. The slopes for the
4 TEOM-FRM, PCM-FRM, and RPM-FRM correlations were 1.01, 0.94, and 0.91, respectively;
5 whereas the y-intercepts for each were 0.68, 0.04, and 0.98. Particulate sulfate measurements on
6 the FRM Teflon filter, the PCM Teflon filter, and PCM Nylon filter were nearly identical.
7 Nitrate results from the three filters were much less consistent, with the FRM collecting
8 substantially less nitrate than that collected on either the denuded nylon filter or a denuder
9 followed by a Teflon-nylon filter sandwich. Particulate ammonia measurements were also
10 compared and showed more scatter than the sulfate measurements but less than the nitrate
11 measurements.
12 An intercomparison of both PM10 and PM2 5 mass measurements was conducted during the
13 1998 Baltimore PM Study (Williams et al., 2000). PM monitors were collocated at a residential
14 indoor, residential outdoor, and ambient monitoring site within Baltimore County, MD. PM
15 samplers included TEOMs, PM25 FRMs, cyclone-based inlets manufactured by University
16 Research Glassware (URG), and Versatile Air Pollution Samplers (VAPS). The VAPS sampler
17 is a dichotomous sampler operating at 33 L/min (one coarse particle channel at 3 L/min, and two
18 fine particle channels at 15 L/min, each). In the configuration employed during this study, one
19 fine particle channel was operated with a Teflon filter backed by a nylon filter and preceded by a
20 sodium carbonate coated annular denuder; the second fine particle channel had a quartz filter
21 preceded by a citric acid-coated annular denuder; and the coarse particle channel had a
22 polycarbonate filter followed by a Zefluor filter for flow distribution. Differences in PM2 5 mass
23 concentrations between the samplers, although not large, were attributed to potential particle
24 nitrate losses, denuder losses, and losses of SVOC for some samplers. Differences between
25 coarse particulate mass concentrations, on the other hand, varied widely between the instruments.
26 In another intercomparison study, Tolocka et al. (200Ib) examined the magnitude of
27 potential sampling artifacts associated with the use of the FRM by collocating FRMs alongside
28 other chemical speciation samplers at four U.S. cities. The locations included a high nitrate and
29 carbon, low sulfate site (Rubidoux, CA); high crustal, moderate carbon and nitrate site
30 (Phoenix); high sulfate, moderate carbon, and low nitrate (Philadelphia); and low PM25 mass
31 (Research Triangle Park, NC). The use of Teflon and heat-treated quartz filters was also
April 2002 2-72 DRAFT-DO NOT QUOTE OR CITE
-------
1 examined in this study. The Teflon filters collected less nitrate than the heat-treated quartz
2 filters. Filters in samplers using denuders to remove organic gases collected less organic PM
3 than filters in samplers without denuders.
4 Peters et al. (2001b) compiled the results of several field studies in which a number of
5 FRM and other PM2 5 samplers were intercompared. In addition to the FRM samplers listed in
6 Table 2-3, other PM2 5 samplers included the Sierra Instruments dichotomous sampler, the
7 Harvard impactor, the IMPROVE sampler, and the Air Metrics saturation monitor. Results were
8 compiled from PM2 5 field studies conducted in Birmingham, Denver, Bakersfield, Phoenix,
9 Research Triangle Park, Atlanta, and Rubidoux. Limited studies on precision for the non-FRM
10 samplers suggest that the Harvard Impactor and dichotomous samplers had the lowest coefficient
11 of variations (CV), with both under 10%. The CV for this study was calculated by dividing the
12 sample standard deviation by the average concentration. The IMPROVE samplers had CVs
13 between 10 and 12%, and the Air Metrics samplers had the highest observed CVs, over 15%.
14 In intercomparisons with FRM samplers, the non-FRM samplers showed strong linear
15 relationships in comparison to the FRM sampler; however, none of the comparisons passed the
16 current EPA Subpart C equivalent method criteria, which EPA is in the process of revising.
17 Detailed information on precision of PM samplers used in monitoring networks may be
18 found in EPA's Technology Transfer Network website (U.S. Environmental Protection Agency,
19 2002).
20
21 2.2.4.1.3 PM1025
22 Measurement techniques for PM10_2 5 are somewhat more complex than those for PM2 5 or
23 PM10 because, for PM10_2 5, it is necessary to isolate a size fraction between an upper 50% cut
24 point of 10 //m Da and a lower 50% cut point of 2.5 //m Da.
25
26 The Difference Method. One approach to measurement of PM10_2 5 is to make separate
27 measurements of PM10 and PM2 5 and take the difference of the resulting equilibrated masses.
28 One problem is that, if either the PM2 5 or the PM10 sampler fails, no PM10_2 5 measurement can be
29 obtained. In addition, errors in cut-point, flow rate, and filter weights (both before use and after
30 collection and equilibration of particles) and uncertainties due to loss of semivolatile components
31 of PM may occur for each size cut. Careful control of flow rate and equivalent treatment of PM10
April 2002 2-73 DRAFT-DO NOT QUOTE OR CITE
-------
1 and PM25 filters in terms of pressure drop across the filter and temperature of the filter during
2 and after collection can improve precision and accuracy. Allen et al. (1999b) summarized
3 several sampling issues to consider in measuring coarse particulate mass by difference, including
4 the use of identical instrumentation (except cutpoints), filter media, filter face velocity, and
5 ambient-filter temperature differences; common flow measurement devices; use of higher
6 sampler flow rates (10 L/min minimum for 24-h sample is recommended); and avoiding
7 excessive filter loading. The concern, expressed by Allen et al. (1999b), that the "pie-plate" inlet
8 required by the final version of the PM2 5 FRM might have a different cut point than the "flat-
9 top" inlet of the PM10 FRM, has been addressed by a wind tunnel study which found both to have
10 an appropriate PM10 cut point (Tolocka et al., 200la).
11 Since the difference method requires weighing two filters, the key to obtaining high
12 precision in the coarse mass measurement is precise measurements of filter weights. Allen et al.
13 (1999b) discuss techniques for increasing the precision of the difference method by reducing
14 errors in filter weights. These include proper temperature and humidity controls, use of a high
15 quality microbalance, 100% replicate weighings, control of static charge, aging of new filters,
16 weighing of a sufficient number of laboratory blank filters, and accounting for buoyancy errors
17 caused by variability in barometric pressure. Allen et al. (1999b) emphasize the necessity of
18 replicate weighing of filters and a third weighing if the difference between the first two weights
19 exceeds a specified minimum. Lawless et al. (2001) investigated the magnitude of uncertainties
20 attributed to fluctuations in some of these parameters (humidity, temperature, drafts, vibration,
21 and electrostatic charges) and recommended methods for improving their control. Koistinen
22 et al. (1999) and Hanninen et al. (2002) give a excellent discussion of the procedures developed
23 to overcome problems associated with gravimetric measurements of PM25 mass in the EXPOLIS
24 (The Air Pollution Exposure Distributions Within Adult Urban Populations in Europe) Study.
25 They discuss factors such as corrections for buoyancy, elimination of static charge, and increases
26 in the mass of blank filters with time. The establishment of a temperature and humidity
27 controlled room required for the equilibration and weighing of filters for the FRM is expensive.
28 Allen et al. (2001) describe a more cost-effective technique that uses a chamber with relative
29 humidity controlled at 34% relative humidity by a saturated aqueous solution of MgCl2.
30 Allen et al. (1999b) recommend that, in reporting precision from collocated samplers both
31 the (CV) and the square of the correlation coefficient (r2) be reported. For a study in Boston with
April 2002 2-74 DRAFT-DO NOT QUOTE OR CITE
-------
1 27 pairs of mass data from collocated PM10 and PM25 using standard weighing methods, they
2 obtained a CV of 4.7% and an r2 of 0.991 for PM25, a CV of 4.4% and an r2 of 0.994 for PM10,
3 and a CV of 15% and an r2 of 0.88 for PM10_25. By using duplicate weighings and other
4 techniques suggested for improving precision, they obtained a CV of 1.3% and an r2 of 0.998 for
5 PM25 in a study in Chicago with 38 collocated measurements. On the basis of the improvement
6 in the CV for PM2 5, they estimate that use of the recommended techniques for PM10_2 5 by
7 difference would have yielded a CV of 3.8% and an r2 of 0.98 if they had been applied in the
8 Chicago study.
9 This "difference" technique has been used to measure PM10_2 5 in a number of studies.
10 Currently, estimates of PM10_25 are obtained by subtracting PM25 from PM10 (both measured by
11 FRM monitors). EPA is currently in the process of developing an FRM for PM10_2 5.
12
13 Multistage Impaction. A second technique involves the use of impaction to isolate the size
14 fraction between 2.5 and 10 //m Da. In the impaction process, the air stream is first accelerated
15 through a small hole (nozzle) or slit. The air stream is directed so that it "impacts" on a surface.
16 Depending on the velocity and pressure of the air stream, particles smaller than a certain size will
17 follow the air stream around the impactor surface. Larger particles will impact on the surface.
18 In practice, impactors have 50% cut points similar to those for the rejection of larger particles in
19 PM25 and PM10 samples (Figure 2-6).
20 Multistage impactors are used to separate particles into several size fractions for the
21 determination of mass and chemical composition as a function of size (Wang and John, 1988;
22 Marple et al., 1991). The major problem with the use of impactors to separate the 10-2.5 //m Da
23 fraction of coarse particles (thoracic coarse PM) is bounce. Coarse particles tend to be dry, solid
24 particles. When they hit a hard surface, they can bounce and be carried away with the air stream
25 (e.g., Dzubay et al., 1976; Wesolowski et al., 1977; Rao and Whitby, 1978; Cheng and Yeh,
26 1979; Wang and John, 1987; John and Sethi, 1993). Various techniques have been used to
27 reduce bounce. One technique is to use a porous substance such as a glass or quartz fiber filter
28 (Chang et al., 1999) material or a polyurethane foam (Breum, 2000; Kavouras and Koutrakis,
29 2001). These techniques may result in less precise separation and yield a sample that must be
30 extracted before chemical analyses can be performed. Another technique is to coat the impactor
31 with a soft wax or grease (Rao and Whitby, 1977; Turner and Hering, 1987; Pak et al., 1992).
April 2002 2-75 DRAFT-DO NOT QUOTE OR CITE
-------
1 This can cause problems with weighing and chemical analyses. In addition, as the deposit of
2 particles builds up, incoming particles may not hit the soft surface, but instead hit a previously
3 collected hard particle and bounce off of it. The WINS impactor discussed earlier uses a filter in
4 a well of low volatility oil to ensure a wetted surface at all times. However, such a technique,
5 while appropriate for removing unwanted particles, would not yield a particle sample suitable for
6 weighing or for chemical analyses.
7
8 Virtual Impaction. In an effort to alleviate the bounce problem, aerosol scientists have
9 developed the "virtual" impactor (Loo et al., 1976; Jaklevic et al., 1977; Loo and Cork, 1988).
10 A hole is placed in the impaction plate just below the accelerating jet. Two controlled flows
11 allow a fraction, e.g., 10% (or another predetermined fraction, typically 5 - 20%), of the air to go
12 through the hole and through a filter (minor flow). A 10% minor flow gives a coarse channel
13 enrichment factor of 10. The remaining fraction (e.g., 90% of the airflow) follows a different
14 path and goes through a second filter (major flow). The upper cutpoint is usually set by the inlet
15 (e.g., 10 //m Da). The flow rates, pressures, and distance from the nozzle to the virtual impactor
16 surface can be varied to direct particles with an Da greater than the lower cutpoint (i.e., > 2.5 //m)
17 to go through the hole and be collected on the first filter and to direct smaller particles (i.e.,
18 < 2.5 //m) to flow around the impactor be collected on the second filter (Marple and Chien,
19 1980). This technique overcomes the problem of bounce. However, a fraction of the smaller
20 particles, equal to the minor flow, will go through the virtual impaction opening with the air
21 stream and be collected on the course particle filter. Thus, in order to determine the mass or
22 composition of the coarse particles, it is necessary to determine the mass and composition of the
23 fine particles and subtract the appropriate fraction from the mass or composition of the particles
24 collected on the coarse particle filter. Virtual impactors that separate particles into two size
25 fractions are known as dichotomous samples. Allen et al. (1999b) discuss potential errors in the
26 dichotomous sampler caused by uncertainties in the coarse mass channel enrichment factor.
27 An example of the separation into fine and coarse particles is shown in Figure 2-17.
28 The dichotomous sampler was developed for use in the Regional Air Monitoring Study
29 (RAMS), part of the Regional Air Pollution Study (RAPS), conducted in St. Louis, Missouri in
30 the mid-1970s (Loo et al., 1976). Dichotomous samplers were a new concept at that time, and
31 there was concern that particle loses might be high at cut point sizes below 2.5 //m Da.
April 2002 2-76 DRAFT-DO NOT QUOTE OR CITE
-------
0 5 10
i i i i I i i i i I
Centimeters
Coarse Particle
Filter
Fine Particle
Filter
Figure 2-17. Schematic diagram showing the principle of virtual impaction. The initial
flow, Qo, is split into a minor flow, Qt, which carries the larger particles,
which impact into the hole, to the coarse particle filter and a major flow, Q2,
which carries the smaller particles, which can follow the airflow, to the fine
particle filter (Loo et al., 1976).
1
2
3
4
5
6
7
In subsequent years, the theory of virtual impaction has advanced. Now virtual impactors, with
rectangular slits or round holes, are used to give cut point sizes as low as 0.15 //m Da and are
used to concentrate coarse, accumulation, and ultrafine mode particles for use in health studies
(Solomon et al., 1983; Marple et al., 1990; Sioutas et al., 1994b,c,d). Dichotomous samplers
were also used in a national network to measure PM2 5 and PM10_2 5 in the Harvard Six City Study
(Dockery et al., 1993) and the Inhalable Particulate Network (Suggs and Burton, 1983).
April 2002
2-77
DRAFT-DO NOT QUOTE OR CITE
-------
1 2.2.5 Speciation Monitoring
2 Speciation Network and Monitoring
3 In addition to FRM sampling to determine compliance with PM standards, EPA requires
4 states to conduct chemical Speciation sampling primarily to determine source categories and
5 trends (Code of Federal Regulations, 200 Ib). Source category apportionment calculations are
6 discussed in Chapter 3. A PM25 chemical speciation network has been deployed that consists of
7 54 core National Ambient Monitoring Stations (NAMS) and approximately 250 State and Local
8 Air Monitoring Stations (SLAMS). In addition, over 100 IMPROVE (Interagency Monitoring of
9 Protected Visual Environments) samplers located at regional background and transport sites can
10 be used to fulfill SLAMS requirements. The overall goal of the speciation program is "to
11 provide ambient data that support the Nation's air quality program objectives" (U.S.
12 Environmental Protection Agency, 1999). Information and reports on EPA's speciation
13 monitoring program may be found on EPA's Technology Transfer Network at
14 http://www.epa.gov/ttn/amtic/pmspec.html. The NAMS speciation sites will provide routine
15 chemical speciation data that will be used to develop annual and seasonal aerosol
16 characterization, air quality trends analysis, and emission control strategies. The SLAMS
17 speciation sites will further support the NAMS network and provide information for
18 development of State Implementation Plans (SIPs).
19 At both NAMs and SLAMs sites, aerosol samples will be collected for analysis of trace
20 elements, ions (sulfate, nitrate, ammonium, sodium, and potassium), and total carbon. The
21 NAMS speciation sites will operate on a 1 in 3 day schedule, with 10 of these sites augmented
22 with continuous speciation analyses for everyday operation. The SLAMS speciation sites will
23 generally operate on a 1 in 6 day basis; however, many sites may be operated on a 1 in 3 day
24 basis in locations where increased data collection is needed. The current samplers include three
25 filters: (1) Teflon for equilibrated mass and elemental analysis by energy dispersive x-ray
26 fluorescence (EDXRF), (2) a nitric acid denuded Nylon filter for ion analysis (ion
27 chromatography), (3) a quartz fiber filter for elemental and organic carbon (but currently without
28 any correction for positive or negative artifacts caused by adsorption of organic gases or the
29 quartz filters or evaporation of semivolatile organic compounds from the collected particles); and
30 (4) thermal optical analysis via NIOSH (National Institute for Occupational Safety and Health)
31 method 5040 (Thermal Optical Transmission) [TOT]). There are several samplers that are
April 2002 2-78 DRAFT-DO NOT QUOTE OR CITE
-------
1 suitable for use in the NAMS/SLAMS network. These samples include an inlet outpoint
2 comparable to the WINS FRM; proven denuder technology for ions; and sampler face velocity
3 and sample volume similar to that of the FRM with 46.2-mm diameter filters.
4 Since 1987, the IMPROVE network has provided measurements of ambient PM and
5 associated light extinction in order to quantify PM chemical components that affect visibility at
6 Federal Class 1 areas that include designated national parks, national monuments, and wilderness
7 areas. Management of this network is a cooperative effort between U.S. EPA, federal land
8 management agencies, and state governments. The IMPROVE program has established
9 protocols for analysis of aerosol measurements that provide ambient concentrations for PM10,
10 PM25, sulfates, nitrates, organic and elemental carbon, crustal material, and a number of other
11 elements. Information on the IMPROVE program may be found at http://vista.cira.colostate.edu/
12 improve.
13 IMPROVE aerosol monitoring consists of a combination of particle sampling and sample
14 analysis. The IMPROVE sampler, which collects two 24-hour duration samples per week,
15 simultaneously collects one sample of PM10 on a Teflon filter, and three samples of PM25 on
16 Teflon, nylon, and quartz filters. PM10 mass concentrations are determined gravimetrically from
17 the PM10 filter sample, while PM25 mass concentrations are determined gravimetrically from the
18 PM25 Teflon filter sample. The PM25 Teflon filter sample is also used to determine
19 concentrations of selected elements using particle-induced x-ray emission (PIXE), x-ray
20 fluorescence (XRF), and Proton Elastic Scattering Analysis (PESA). The PM2 5 nylon filter
21 sample, which is preceded by a denuder to remove acidic gases, is analyzed to determine nitrate
22 and sulfate aerosol concentrations using Ion Chromatography (1C). Finally, the PM2 5 quartz
23 filter sample is analyzed for organic and elemental carbon using the Thermal Optical Reflectance
24 (TOR) method.
25 Several of the PM2 5 size selectors developed for use in the EPA National PM25 Chemical
26 Speciation Trends network were recently evaluated by comparing their penetration curves under
27 clean room experiments with that of the WINS impactor (Peters et al., 200 Ic). The
28 corresponding speciation monitors were then compared to the FRM in four cities. The PM2 5
29 inlets tested were the SCC 2.141 cyclone (6.7 L/min) that is in the Met One Instruments SASS
30 sampler; the SCC 1.829 cyclone (5.0 L/min) that is proposed for use in the Rupprecht and
31 Patashnik real-time sulfate/nitrate monitor; the AN 3.68 cyclone (24.0 L/min) that is in the
April 2002 2-79 DRAFT-DO NOT QUOTE OR CITE
-------
1 Andersen RAAS; and the spiral separator (7.0 Lpm) that was previously in the Met One SASS.
2 The outpoints of the SCC cyclones compared reasonably well with the WINS (2.52 and
3 2.44 micrometers for the SCC 2.141 and SCC 1.829, respectively, at their design flowrates), but
4 both demonstrated a tail extending into the coarse particle mode. The AN inlet had the sharpest
5 cutpoint curve, but the 50% cutpoint diameter was 2.7 //m Da at its design flowrate. The spiral
6 inlet had the shallowest cutpoint curve, and the 50% cut point was 2.69 and 2.67 //m Da for an
7 ungreased and greased inlet, respectively. The speciation samplers were also compared to the
8 FRM sampler with WINS inlet under ambient conditions in four cities. The Andersen RAAS
9 equipped with the AN 3.68 cyclone compared well to the FRM in all four cities, when compared
10 on the basis of PM2 5 mass, sulfate, and crustal concentrations. Greasing the spiral inlet in the
11 Met One sampler improved the performance of that sampler, which tended to give much higher
12 PM25 concentrations than the FRM in cities with high crustal particulate matter.
13
14 2.2.6 Inorganic Elemental Analyses
15 In addition to the lighter elements, hydrogen, carbon, oxygen and nitrogen, the following
16 40 heavier elements are commonly found in ambient air samples: sodium, magnesium,
17 aluminum, silicon, phosphorus, sulfur, chlorine, potassium, calcium, titanium, vanadium,
18 chromium, manganese, iron, cobalt, nickel, copper, zinc, gallium, arsenic, selenium, bromine,
19 rubidium, strontium, yttrium, zirconium, molybdenum, palladium, silver, cadmium, indium, tin,
20 antimony, barium, lanthanum, gold, mercury, thallium, lead, and uranium. These often indicate
21 air pollution sources and several of them are considered to be toxic (transition metals,
22 water-soluble metals, and metals in certain valence states [e.g., Fe(II), Fe(ni), Cr(in), Cr(VI),
23 As(ni), As(V)]). Measurement methods for the heavier elements include: (1) energy dispersive
24 x-ray fluorescence (EDXRF); (2) synchrotron induced X-ray emission (S-XRF); (3) proton
25 induced x-ray emission (PIXE); (4) proton elastic scattering analysis (PESA); (5) total reflection
26 X-ray fluorescence (TRXRF); (6) instrumental neutron activation analysis (INAA); (7) atomic
27 absorption spectrophotometry (AAS); (8) inductively coupled plasma with atomic emission
28 spectroscopy (ICP-AES); (9) inductively coupled plasma with mass spectroscopy (ICP-MS); and
29 (10) scanning electron microscopy (SEM). These methods differ with respect to detection limits,
30 sample preparation, and cost (Chow, 1995). XRF and PIXE are the most commonly applied
31 methods because they quantify more than 40 detectable elements, they are non-destructive, and
April 2002 2-80 DRAFT-DO NOT QUOTE OR CITE
-------
1 they are relatively inexpensive. Both were discussed in the previous 1996PM AQCD. TRXRF
2 and S-XRF are newer techniques with lower detection limits. AAS, ICP-AES, and ICP-MS are
3 also appropriate for ionic measurements when the particles are extracted in deionized distilled
4 water. PESA provides a means for measuring elements with lower atomic numbers from
5 hydrogen to carbon.
6
7 2.2.6.1 Energy Dispersive X-ray Fluorescence (EDXRF)
8 EDXRF has usually been the method of choice for analysis of trace elements on filters.
9 EDXRF is preferred for aerosol analysis over wavelength dispersive XRF because it allows fast
10 and simultaneous analysis over the total spectrum, allowing for the analysis of numerous
11 elements simultaneously. EDXRF can accommodate small sample sizes and requires little
12 sample preparation or operator time after the samples are placed into the analyzer. It also leaves
13 the sample intact after analysis; so, further analysis is possible. XRF irradiates a uniform particle
14 deposit on the surface of a membrane filter with 1 to 50 kev x-rays that eject inner shell electrons
15 from the atoms of each element in the sample (Dzubay and Stevens, 1975; Jaklevic et al., 1977;
16 Billiet et al., 1980; Potts and Webb, 1992; Piorek, 1994; Bacon et al., 1995; deBoer et al., 1995;
17 Holynska et al., 1997; Torok et al., 1998; Watson et al., 1999). When a higher energy electron
18 drops into the vacant lower energy orbital, a fluorescent x-ray photon is released. The energy of
19 this photon is unique to each element, and the number of photons is proportional to the
20 concentration of the element. Concentrations are quantified by comparing photon counts for a
21 sample with those obtained from thin-film standards of known concentration (Dane et al., 1996).
22 The previous 1996 PM AQCD included a detailed discussion of EDXRF.
23 Emitted x-rays with energies less than ~4 kev (affecting the elements sodium, magnesium,
24 aluminum, silicon, phosphorus, sulfur, chlorine, and potassium) are absorbed in the filter, in a
25 thick particle deposit, or even by large particles in which these elements are contained. Very
26 thick filters also scatter much of the excitation radiation or protons, thereby lowering the
27 signal-to-noise ratio for XRF and PFXE. For this reason, thin membrane filters with deposits in
28 the range of 10 to 50 //g/cm2 provide the best accuracy and precision for XRF and PIXE analysis
29 (Davis et al., 1977; Haupt et al., 1995).
30
31
April 2002 2-81 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.2.6.2 Synchrotron Induced X-ray Fluorescence (S-XRF)
2 S-XRF is a form of EDXRF in which the exciting x-rays are derived from a synchrotron.
3 Bremmstrahlung x-rays are generated when energetic electrons (generally several GeV in energy)
4 are forced by a magnetic field to make a bend in their path. The advantages of the technique are
5 that an extremely high flux of x-rays may be obtained and that the x-rays are 100% polarized in
6 the plane of the electron beam. The former allows for x-ray beams generally of 50 to 500 //m in
7 diameter. However, the beams can be focused into x-ray microprobes, with spot sizes on the
8 order of one //m diameter. The x-ray polarization allows for removal of most of the background
9 normally found under the characteristic x-ray peaks, greatly improving sensitivity compared to
10 other XRF techniques. The primary disadvantages are the limited number of synchrotrons, and
11 that few synchrotrons have S-XRF capabilities. Thus, the technique has been relatively little
12 used for PM, and then generally for special problems such as the smoke from the Kuwaiti oil
13 fires (Cahill et al., 1992, Reid et al., 1994). However, with the increasing availability of S-XRF
14 facilities dedicated to PM analysis, the first of which was the Advanced Light Source opened at
15 Lawrence Berkeley National Laboratory last year, utilization of S-XRF for PM analysis is
16 increasing.
17
18 2.2.6.3 Proton Induced X-ray Emission (PIXE)
19 PFXE differs from XRF analysis in the excitation source for producing fluorescence. The
20 filter deposit is bombarded with high-energy protons to remove inner shell electrons and the
21 resulting characteristic x-rays are analyzed as in XRF (Johansson, 1970, Cahill, 1981, 1985;
22 Zeng et al., 1993). Small accelerators, generally Van de Graaffs, generate intense beams of low
23 energy protons, generally of a few MeV in energy. These have the ability to remove electrons
24 from inner shells of atoms of any element. Thus, PFXE can see a very wide range of elements in
25 a single analysis. The cross section for producing x-rays using protons of a few MeV in energy
26 tends to favor lighter elements, Na through Ca, but sensitivities for equivalent PFXE and
27 multi-wavelength XRF analysis are roughly comparable. The technique has been widely used in
28 the U.S. (Flocchini et al., 1976, Malm et al., 1994) and around the world, as many universities
29 have the small accelerators needed for the method. Like S-XRF, the proton beams can be
30 focused into //m size beams, but these have been relatively little used for PM. However, the mm
April 2002 2-82 DRAFT-DO NOT QUOTE OR CITE
-------
1 size beams used in both S-XRF and PIXE are well suited to analyzing the limited mass and small
2 deposits that result from detailed particle size measurements by impactors. (Perry et al., 1999)
3
4 2.2.6.4 Proton Elastic Scattering Analysis (PESA)
5 With the routine availability of elemental analyses for all elements sodium and heavier,
6 organic components remain the major unmeasured species for mass balance. For programs like
7 IMPROVE (Malm et al., 1994), parallel filters are collected for separate organic and elemental
8 carbon determinations. Aerosol programs that use PIXE can directly measure hydrogen
9 simultaneously by scattering protons from Teflon filters that lack hydrogen (Cahill et al., 1989,
10 1992). Generally, results from organic matter by carbon combustion from quartz filters and
11 organic matter by hydrogen from Teflon filters are in agreement, assuming certain assumptions
12 about the chemical states of sulfates and nitrates are met (Malm et al., 1994, Cahill et al., 1996).
13
14 2.2.6.5 Total Reflection X-ray Fluorescence (TRXRF)
15 One of the limitations of the EDXRF method is the minimum detection limit, which may
16 be high due to high background values (Streit et al., 2000). By implementation of x-ray optical
17 geometries that use the total reflection of the primary radiation on flat surfaces, scattering on the
18 substrate is reduced, so that detection limits can be reduced. This is the basis for the total
19 reflection x-ray fluorescence (TRXRF) method (Aiginger and Streli, 1997). This modification to
20 the EDXRF technique improves detection limits and avoids the need to correct for matrix effects.
21 Despite its apparent advantages, TRXRF has not yet become widely in use for atmospheric
22 aerosol analysis but has been used in the analysis of marine aerosol (Stahlschmidt et al., 1997)
23 and at a high elevation site (Streit et al., 2000). Streit et al. sampled ambient air at the High
24 Alpine Research Station (3580 m above sealevel) in the Bernese Alps, Switzerland, using a nine-
25 stage, single-jet, low-pressure, cascade impactor equipped with quartz impactor plates coated
26 with silicon oil diluted in 2-propanol. The typical sample volume for a weekly sample was
27 10 m3. The quartz plates were analyzed directly by TRXRF. Streit et al. reported that the
28 minimum detection limits, defined by the 3 a values of the blanks, ranged from 25 ng for S,
29 decreased monotonically with increasing atomic number down to 5 pg for Rb, and decreased
30 after that. The use of TRXRF is expected to increase as EDXRF users become aware of the
April 2002 2-83 DRAFT-DO NOT QUOTE OR CITE
-------
1 method. A relatively low-cost, add-on unit has been developed that would allow EDXRF users
2 to test the TRXRF technique (Aiginger, 1991).
3
4 2.2.6.6 Instrumental Neutron Activation Analysis (INAA)
5 INAA irradiates a sample in the core of a nuclear reactor for few minutes to several hours,
6 depending on the elements being quantified (Dams et al., 1970; Zoller and Gordon, 1970;
7 Nadkarni, 1975; Landsberger, 1988; Olmez, 1989; Ondov and Divita, 1993). The neutron
8 bombardment chemically transforms many elements into radioactive isotopes. The energies of
9 the gamma rays emitted by these isotopes identify them and, therefore, their parent elements.
10 The intensity of these gamma rays is proportional to the amount of the parent element present in
11 the sample. Different irradiation times and cooling periods are used before counting with a
12 germanium detector. In source apportionment studies, it is possible to use a combination of XRF
13 and INAA to develop a relatively complete set of elemental measurements. Between these two
14 analytical techniques, good sensitivity is possible for many elements, including most of the toxic
15 metals of interest. In general, XRF provides better sensitivity for some metals (e.g., Ni, Pb, Cu,
16 and Fe); whereas INAA provides better sensitivity for others (Sb, As, Cr, Co, Se, and Cd). Both
17 methods provide similar detection limits for still other elements (V, Zn, and Mn). INAA does
18 not quantify some of the abundant species in ambient particulate matter such as silicon, nickel,
19 tin, and lead. While INAA is technically nondestructive, sample preparation involves folding the
20 sample tightly and sealing it in plastic, and the irradiation process makes the filter membrane
21 brittle and radioactive. These factors limit the use of the sample for subsequent analyses.
22 INAA has been used to examine the chemical composition of atmospheric aerosols in
23 several studies either as the only method of analysis or in addition to XRF (e.g., Yatin et al.,
24 1994; Gallorini, 1995). INAA has higher sensitivity for many trace species, and it is particularly
25 useful in analyzing for many trace metals. Landsberger and Wu (1993) analyzed air samples
26 collected near Lake Ontario for Sb, As, Cd, In, I, Mo, Si, and V using INAA. They demonstrated
27 that using INAA in conjunction with epithermal neutrons and Compton suppression produces
28 very precise values with relatively low detection limits.
29 Enriched rare-earth isotopes have been analyzed via INAA and used to trace sources of
30 particulate matter from a coal-fired power plant (Ondov et al., 1992), from various sources in the
31 San Joaquin Valley (Ondov, 1996), from intentially tagged (iridium) diesel emissions from
April 2002 2-84 DRAFT-DO NOT QUOTE OR CITE
-------
1 sanitation trucks (Suarez et al., 1996; Wu et al., 1998), and from iridium-tagged emissions from
2 school buses (Wu et al., 1998).
3 An intercomparison was conducted in which 18 pairs of filters were sent to participants in
4 the Coordinated Research Program (CRP) on Applied Research on Waste Using Nuclear Related
5 Analytical Techniques (Landsberger et al., 1997). As part of that study, participants used PIXE,
6 INAA, XRF, or AAS to analyze the samples. Many of the results for XRF and PIXE in the
7 coarse fraction were observed to be biased low compared to INAA. The authors speculated that
8 there is a systematic error because of self-attenuation of the x-rays resulting from the particle size
9 effect.
10
11 2.2.6.7 Atomic Absorption Spectrophotometry (AAS)
12 AAS is applied to the residue of a filter extracted in a strong solvent to dissolve the solid
13 material; the filter or a portion of it is also dissolved during this process (Ranweiler and Moyers,
14 1974; Fernandez, 1989; Jackson and Mahmood, 1994; Chow et al., 2000a). A few milliliters of
15 this extract are injected into a flame where the elements are vaporized. Elements absorb light at
16 certain wavelengths in the visible spectrum, and a light beam with wavelengths specific to the
17 elements being measured is directed through the flame to be detected by a monochrometer. The
18 light absorbed by the flame containing the extract is compared with the absorption from known
19 standards to quantify the elemental concentrations. AAS requires an individual analysis for each
20 element, and a large filter or several filters are needed to obtain concentrations for a large variety
21 of elements. AAS is a useful complement to other methods, such as XRF and PIXE, for species
22 such as beryllium, sodium, and magnesium which are not well-quantified by these methods.
23 Airborne particles are chemically complex and do not dissolve easily into complete solution,
24 regardless of the strength of the solvent. There is always a possibility that insoluble residues are
25 left behind and that soluble species may co-precipitate on them or on container walls.
26 AAS was used to characterize the atmospheric deposition of trace elements Zn, Ni, Cr, Cd,
27 Pb, and Hg to the Rouge River watershed by particulate deposition (Pirrone and Keeler, 1996).
28 The modeled deposition rates were compared to annual emissions of trace elements that were
29 estimated from the emissions inventory for coal and oil combustion utilities, iron and steel
30 manufacturing, metal production, cement manufacturing, and solid waste and sewage sludge
31 incinerators. They found generally good agreement between the trend observed in atmospheric
April 2002 2-85 DRAFT-DO NOT QUOTE OR CITE
-------
1 inputs to the river (dry + wet deposition) and annual emissions of trace elements, with R2s
2 varying from «0.84 to 0.98. Both atmospheric inputs and emissions were found to have followed
3 downward trends for Pb. For the period of 1987 to 1992, steady increases were observed for Cd
4 (major sources are municipal solid waste incineration, coal combustion, sludge incineration, and
5 iron and steel manufacturing), Cr and Ni (major sources are iron and steel production and coal
6 combustion), and Hg (major sources are coal, the contribution from which had decreased from
7 53 to 45%, and municipal, solid, and medical waste incineration, the contribution from which has
8 increased).
9
10 2.2.6.8 Inductively Coupled Plasma with Atomic Emission Spectroscopy (ICP-AES)
11 ICP-AES introduces an extracted sample into an atmosphere of argon gas seeded with free
12 electrons induced by high voltage from a surrounding Tesla coil (Fassel and Kniseley, 1974;
13 McQuaker et al., 1979; Lynch et al., 1980; Harman, 1989; Tyler, 1992; Baldwin et al., 1994).
14 The high temperatures in the induced plasma raise valence electrons above their normally stable
15 states. When these electrons return to their stable states, a photon of light is emitted that is
16 unique to the element that was excited. This light is detected at specified wavelengths to identify
17 the elements in the sample. ICP-AES acquires a large number of elemental concentrations using
18 small sample volumes with acceptable detection limits for atmospheric samples. As with AAS,
19 this method requires complete extraction and destruction of the sample.
20
21 2.2.6.9 Inductively Coupled Plasma with Mass Spectroscopy (ICP-MS)
22 ICP-MS has been applied in the analysis of personal exposure samples (Tan and Horlick,
23 1986; Gray and Williams, 1987a,b; Nam et al., 1993; Munksgaard and Parry, 1998; Campbell
24 and Humayun, 1999). Ion species generated from ICP and from the sample matrix can produce a
25 significant background at certain masses resulting in formation of polyatomic ions that can limit
26 the ability of ICP-MS to determine some elements of interest. Cool plasma techniques have
27 demonstrated the potential to detect elements at the ultra-trace level (Nham et al., 1996) and to
28 minimize common molecular ion interferences (Sakata and Kawabata, 1994; Turner, 1994;
29 Plantz, 1996). Detection limits of ICP-MS using a one-second scan are typically in the range of
30 10"3 ng/m3, which is an order of magnitude lower than other elemental analysis methods. The
31 instrument can also be set up to analyze a wide dynamic range of aerosol concentrations. Isotope
April 2002 2-86 DRAFT-DO NOT QUOTE OR CITE
-------
1 analysis can also be performed with ICP-MS. Intercomparison studies are needed to establish the
2 comparability of ICP-MS with other non-destructive filter analysis methods.
3 Keeler and Pirrone (1996) used ICP-MS to determine trace elements Cd, Mn, V, As, Se,
4 and Pb in atmospheric fine particulate (PM2 5) and total suspended particulate samples collected
5 in two Detroit sites. The results were used in a deposition model with additional measurements
6 using AAS to estimate the dry deposition flux of trace elements to Lake Erie.
7
8 2.2.6.10 Scanning Electron Microscopy (SEM)
9 Mamane et al. (2001) investigated the use of computer-controlled scanning electron
10 microscopy (CCSEM) as a way of supplementing XRF analysis and providing automated
11 analysis of particle size, chemistry, and particle classification. An ambient coarse particulate
12 sample from Baltimore was collected on a polycarbonate filter for this analysis. CCSEM
13 analyses were conducted for 2819 particles in 78 randomly selected fields of view during an
14 unattended 8-h run. Mamane et al. confirmed the stability of the CCSEM instrument over
15 several hours of operation. The physical properties of the sample such as particle diameter, mass
16 loading per field, and particle number per field were well represented by analyzing approximately
17 360 particles with little additional information gained by analyzing more particles. Teflon filters
18 are not well suited for SEM analyses. Analysis of fine PM is expected to pose analytical
19 challenges not addressed in the present study (Mamane et al., 2000).
20 Nelson et al. (2000) applied Raman chemical imaging and SEM (Raman/SEM) to study the
21 size, morphology, elemental and molecular composition, and molecular structure of fine
22 particulate matter. In their study, filter compatibility was examined, and Raman/SEM chemical
23 imaging was conducted for several standard materials as well as for ambient PM2 5 samples.
24 Polycarbonate was determined to be a suitable substrate for both SEM and Raman chemical
25 imaging analysis.
26 Conner et al. (2001) used CCSEM with individual X-ray analysis to study the chemical and
27 physical attributes of indoor and outdoor aerosols collected around a retirement home in
28 Baltimore. The CCSEM technique was demonstrated to be capable of identifying spherical
29 particles typical of combustion or other high temperature (presumably industrial) processes as
30 well as pollens and spores. Indoor particles originating from cosmetics were also identified.
31
April 2002 2-87 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.2.7 Elemental and Organic Carbon in Particulate Matter
2 Three classes of carbon are commonly measured in ambient aerosol samples collected on
3 quartz-fiber filters: (1) organic, volatilized, or non-light absorbing carbon (organic carbon, OC);
4 (2) elemental or light-absorbing carbon (elemental carbon, EC); and (3) carbonate carbon (CC).
5 The sum of OC, EC, and CC in PM gives the total carbon (TC). Carbonate carbon (i.e., K2CO3,
6 Na2CO3, MgCO3, CaCO3), which can be determined on a separate filter section by measurement
7 of the carbon dioxide (CO2) evolved upon acidification (Johnson et al., 1980), is usually on the
8 order of 5% or less for particulate samples collected in urban areas (Appel, 1993). Black carbon
9 (BC) refers to an estimate of EC measured by absorption of visible light. The 1996 PM AQCD
10 (U.S. Environmental Protection Agency, 1996a) listed several filter-based, thermal methods for
11 measuring OC and EC and described the thermal/optical reflectance (TOR) method that was
12 noted, along with thermal manganese oxidation, to be one of the most commonly applied
13 methods in the United States at the time. In thermal separation methods, thermally evolved
14 OC and EC are oxidized to CO2 and quantified either by nondispersive infrared detection or
15 electrochemically or by reducing the CO2 to CH4 and quantifying CH4 via flame ionization
16 detection (FID). The various methods give similar results for TC, but not for EC or OC.
17 Chow and Watson (1998) summarize different carbon analysis methods along with their
18 measurement principles. The definitions of organic and elemental carbon are operational (i.e.,
19 method dependent) and reflect the method and purpose of measurement. Elemental carbon is
20 sometimes termed "soot", "graphitic carbon", or "black carbon." For studying visibility
21 reduction, light-absorbing carbon is a more useful concept than elemental carbon. For source
22 apportionment by receptor models, several consistent but distinct fractions of carbon in both
23 source and receptor samples are desired, regardless of their light-absorbing or chemical
24 properties. Differences in ratios of the carbon concentrations in these fractions form part of the
25 source profile that distinguishes the contribution of one source from the contributions of other
26 sources (Watson et al., 1994a).
27 Light-absorbing carbon is not entirely graphitic carbon because there are many organic
28 materials which absorb light (e.g., tar, motor oil, asphalt, coffee). Even the "graphitic" black
29 carbon in the atmosphere has only a poorly developed graphitic structure with abundant surface
30 chemical groups. "Elemental carbon" is a poor but common description of what is measured.
31 For example, a substance of three-bond carbon molecules (e.g., pencil lead) is black and
April 2002 2-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 completely absorbs light, but four-bond carbon in a diamond is completely transparent and
2 absorbs very little light. Both are pure, elemental carbon.
3 Chow et al. (1993) document several variations of the thermal method for organic and
4 elemental carbon. The thermal/optical reflectance (TOR), thermal/optical transmission (TOT),
5 and thermal manganese oxidation (TMO) methods are most commonly used for the analysis of
6 organic and elemental carbon in atmospheric PM. Filter transmission analysis is often performed
7 to estimate particle light absorption which is proportional to the level of elemental carbon in the
8 atmosphere. These methods are discussed in detail in the following subsections.
9 The thermal manganese oxidation (TMO) method (Mueller et al., 1982; Fung, 1990) uses
10 manganese dioxide present and in contact with the sample throughout the analysis as the
11 oxidizing agent. Temperature is relied upon to distinguish between organic and elemental
12 carbon. Carbon evolving at 525 °C is classified as organic carbon, and carbon evolving at
13 850 °C is classified as elemental carbon. This method has been used in the SCENES
14 (Subregional Cooperative Electric Utility, Department of Defense, National Park Services, and
15 Environmental Protection Agency Study); (Sutherland and Bhardwaja, 1987; Mueller et al.,
16 1986) visibility network, as well as in the SCAQS (Southern California Air Quality Study)
17 (Chow et al., 1994a,b; Watson et al., 1993, 1994a,b).
18 The thermal/optical reflectance (TOR) method of carbon analysis developed by Huntzicker
19 et al. (1982) has been adapted by several laboratories for the quantification of organic and
20 elemental carbon in PM collected on quartz-fiber filters. Although the principle used by these
21 laboratories is identical to that of Huntzicker et al. (1982), the details differ with respect to
22 calibration standards, analysis time, temperature ramping, and volatilization/combustion
23 temperatures.
24 In the most commonly applied version of the TOR method (Chow et al., 1993), a filter is
25 submitted to volatilization at temperatures ranging from ambient to 550 °C in a pure helium
26 atmosphere, then to combustion at temperatures between 550 °C to 800 °C in a 2% oxygen and
27 98% helium atmosphere with several temperature ramping steps. The carbon that evolves at each
28 temperature is converted to methane and quantified with a flame ionization detector. The
29 reflectance from the deposit side of the filter punch is monitored throughout the analysis. This
30 reflectance usually decreases during volatilization in the helium atmosphere owing to the
31 pyrolysis of organic material. When oxygen is added, the reflectance increases as the
April 2002 2-89 DRAFT-DO NOT QUOTE OR CITE
-------
1 light-absorbing carbon is combusted and removed. Organic carbon is defined as that which
2 evolves prior to re-attainment of the original reflectance, and elemental carbon is defined as that
3 which evolves after the original reflectance has been attained. By this definition, "organic
4 carbon" is actually organic carbon that does not absorb light at the wavelength (632.8 nm) used,
5 and "elemental carbon" is light-absorbing carbon (Chow et al., 1993).
6 The thermal/optical transmission (TOT) method applies to the same thermal/optical carbon
7 analysis method except that transmission instead of reflectance of the filter punch is measured.
8 The National Institute for Occupational Safety and Health (NIOSH) Method 5040 for monitoring
9 elemental carbon as a marker for particulate diesel exhaust is based upon the TOT method (Birch
10 and Gary, 1996). The TOT OC/EC method consists of a two-stage process with the first stage
11 being conducted in a pure helium atmosphere at temperatures of 250, 500, 650, and 850 °C for a
12 total of 4.5 minutes and the second stage conducted in a 2% oxygen/98% helium mix at
13 temperatures of 650, 750, 850, and 940 °C for 4 minutes. A pyrolysis base correction is made
14 based on the transmission measurement.
15 Chow et al. (1993) document several variations of the thermal (T), thermal/optical
16 reflectance (TOR), thermal/optical transmission (TOT), and thermal manganese oxidation
17 (TMO) methods for organic and elemental carbon. Comparisons among the results of the
18 majority of these methods show that they yield comparable quantities of total carbon in aerosol
19 samples, but the distinctions between organic and elemental carbon are quite different (Cadle and
20 Groblicki, 1982; Cadle and Mulawa, 1990; Countess, 1990; Hering et al., 1990; Birch, 1998;
21 Schmid et al., 2001). TOR was consistently higher than TMO for elemental carbon, especially in
22 woodsmoke-dominated samples where the disparity was as great as sevenfold. For the sum of
23 organic and elemental carbon, these methods reported agreement within 5% to 15% for ambient
24 and source samples (Houck et al., 1989; Kusko et al., 1989; Countess, 1990; Shah and Rau,
25 1990) and within 3% on carefully prepared standards. Evaluation of these methods thus is a
26 matter of assessing how they differentiate between organic and elemental carbon. The TMO
27 method attributes more of the total carbon to organic carbon and less to elemental carbon than
28 the TOR and TOT methods. None of the methods represents an ideal procedure for the
29 separation of organic from elemental carbon.
30 In a methods comparison study (Countess, 1990), it was shown that it is necessary to
31 minimize or correct for pyrolytically generated EC ("char") and that CC found in wood smoke
April 2002 2-90 DRAFT-DO NOT QUOTE OR CITE
-------
1 and automobile exhaust samples may interfere with some of the thermal methods. Recently,
2 Lavanchy et al. (1999) reported on a study in which the operation of a catalytic oxidation system
3 was modified in an attempt to minimize pyrolysis of OC and, at the same time, minimize the
4 contribution of CaCO3. The system uses two ovens, one at 340 °C and one at 650 °C. The filter
5 sample is placed in a moveable sample boat. In order to minimize charring, the sample is first
6 flash heated in the 650 °C oven for 1 min. It is then inserted into the 340 °C stage of the two-
7 stage oven. In both steps, OC is oxidized to CO2 in the presence of O2. After 42 mins, the filter
8 is moved into the second-stage oven. During this third step, EC is oxidized to CO2 at 650°C for
9 32 min. This temperature is reported to be sufficient to completely oxidize EC, but with only
10 about 1% of the CaCO3 being vaporized (Lavanchy et al., 1999; Petzold et al., 1997). To test for
11 charring, they challenged their system with atmospheric samples for which duplicates were
12 analyzed via the German reference method (in which a solvent extraction is used to remove
13 organics before combustion) for measuring OC and EC in atmospheric samples (Petzold and
14 Niessner, 1996). Lavanchy et al. (1999) reported a high correlation (R2 = 0.97) between their
15 thermal oxidation method and the German reference method (VDI). The slope of the EC:EC
16 VDI line was 0.92, and the intercept was -0.37 //g cm"2. They also reported detection limits of
17 1.3 //g for EC and 1.8 //g for OC.
18 Pyrolytic char is corrected for in thermal-optical analysis. In thermal-optical methods
19 (Birch and Gary, 1996; Chow et al., 1993), punches from a quartz sampling filter are inserted
20 into the carbon analyzer and heated in a helium atmosphere to volatilize organic carbon. Then,
21 the temperature is reduced, and oxygen is added to the carrier gas so that desorbed compounds
22 are oxidized to CO2, reduced to methane, and measured in a flame ionization detector. In order
23 to account for the portion of the OC that is pyrolyzed, a He-Ne laser monitors the sample
24 reflectance (or transmittance). As the pyrolysis occurs, the sample gets darker, and the
25 reflectance decreases. As elemental carbon is removed, the filter lightens, and the reflectance
26 increases until all carbon has been removed from the filter. The split between organic and
27 elemental carbon is considered to be the point at which the reflectance regains its prepyrolysis
28 value with material removed prior to this point being considered organic and that after,
29 elemental.
30 The thermal/optical transmission (TOT) method is similar to the TOR with the primary
31 difference being that light transmission rather than reflectance is monitored on the filter
April 2002 2-91 DRAFT-DO NOT QUOTE OR CITE
-------
1 throughout the analysis. The TOT method of Birch and Gary (1996) consists of a two-stage
2 process with the first stage being conducted in a pure helium atmosphere and the second stage
3 conducted in a 10% oxygen-helium mix. The temperature is raised to approximately 820 °C in
4 the helium phase, during which organic and carbonate carbon are volatilized from the filter.
5 In the second stage, the oven temperature is reduced then raised to about 860 °C. During the
6 second stage, pyrolysis correction and EC measurement are made. Figure 2-18, an example of a
7 TOT thermogram, shows temperature, transmittance, and FID response traces. Peaks are evident
8 that correspond to OC, CC, EC, and pyrolitic carbon (PC). As can be seen in this figure, the high
9 temperature in the first stage allows for decomposition of CC. The ability to quantify PC is
10 particularly important in high OC/EC regions (like wood-smoke-impacted airsheds), allowing for
11 the volatilization of any remaining complex organic compounds so they are not apportioned to
12 the EC phase.
13 The NIOSH Method 5040 for monitoring elemental carbon as a marker for particulate
14 diesel exhaust is based upon a TOT method analyzer (Birch and Gary, 1996); whereas the OC/EC
15 method specified for the IMPROVE network is the TOR method (Chow et al., 2000b). Chow
16 et al. (2000b) compared the OC, EC, and TC measurements from NIOSH and IMPROVE
17 methods. The two methods use different temperature and atmospheric controls to separate OC
18 and EC. In addition, the NIOSH (TOT) method uses light transmission through the filter and the
19 IMPROVE (TOR) method uses light reflectance from the filter to measure pyrolyte carbon. The
20 IMPROVE thermal protocol specifies organic carbon fractions at 120, 250, 450, and 550 °C in a
21 nonoxidizing atmosphere (He) and elemental organic fractions at 550, 700, and 800 °C in an
22 oxidizing atmosphere. The NIOSH method differs in its thermal protocol, which has organic
23 carbon fractions at 250, 500, 650, and 850 °C in a nonoxidizing atmosphere (also He) and
24 elemental carbon fractions at 650, 750, and 850 °C in an oxidizing atmosphere. The high
25 temperature before addition of oxygen in the NIOSH method is necessary to quantify particulate
26 carbonate, which evolves between 650 and 830 °C (Birch and Gary, 1996). The two methods
27 also differ in the specified residence times at each temperature setpoint. The residence times at
28 each setpoint are typically longer for the IMPROVE analysis as compared to the NIOSH analysis.
29 Chow et al. (2000b) analyzed 60 quartz filter samples that represented a wide variety of
30 aerosol compositions and concentrations. The TC measurements from each protocol were in
31 good agreement with no statistically significant differences. A statistically significant difference
April 2002 2-92 DRAFT-DO NOT QUOTE OR CITE
-------
OC - EC split
Time (min)
Figure 2-18. This thermogram, for a sample containing rock dust (a carbonate
source) and diesel exhaust, shows three traces that correspond to
temperature, filter transmittance, and FID detector response. Peaks
correspond to organic (OC), carbonate (CC), pyrolytic (PC), and
elemental (EC) carbon. The final peak is a methane calibration peak.
Source: Birch and Gary (1996).
1 was observed in the fraction of TC that is attributed to EC as determined by the IMPROVE and
2 NIOSH thermal evolution protocols with the IMPROVE EC measurements typically higher than
3 the NIOSH EC measurements. This difference was attributed to the 850 °C temperature step in
4 the oxidizing atmosphere in the NIOSH protocol. Chow et al. (2000b) compared the OC for each
5 method and found that the two methods showed good agreement when the 850 °C nonoxidizing
6 temperature step in the NIOSH method was not included in determination of OC. There was also
7 a difference between the reflectance and transmittance detection methods in the pyrolysis
8 adjustment, although this difference was most noticeable for very black filters for which neither
April 2002
2-93
DRAFT-DO NOT QUOTE OR CITE
-------
1 reflectance nor transmittance was able to accurately detect further blackening by pyrolysis.
2 Because OC and EC are operationally defined parameters, Chow et al. (2000b) pointed out that it
3 is important to retain ancillary information when reporting EC and OC by these analytical
4 methods, so that comparisons can be made among measurements taken at different sites using
5 these two methods.
6 Further refinement of thermal techniques has resulted in the evolved gas analysis (EGA)
7 method, described by Grosjean et al. (1994). This technique involves combustion of paniculate
8 matter samples in an oxidizing environment while the temperature is raised from 100 to 600 °C.
9 The amount of evolved CO2 contains information about the volatility of the organic aerosol
10 compounds. Grosjean et al. (1994) present thermograms both for specific organic compounds
11 (e.g., adipic acid) and for specific sources (e.g., vehicular traffic). They suggest that EGA may
12 be useful for source apportionment applications. Kirchstetter et al. (2001) and Novakov et al.
13 (1997) have also used EGA to provide insights regarding organic sampling artifacts.
14 A more recent international intercomparison on the analysis of carbonaceous aerosols on
15 quartz fiber filters was organized by the Vienna University of Technology and involved
16 seventeen laboratories and nine different thermal and optical methods (Schmid et al., 2001).
17 All laboratories were sent punches from three 150-mm quartz fiber filters that had been exposed
18 for 24 h near a high traffic street in Berlin. Five laboratories employed VDI2465 methods that
19 are official methods in Germany. Two of these laboratories used the VDI 2465/1 method that
20 determines extractable organic carbon, non-extractable organic carbon, and elemental carbon.
21 The solvent extraction step incorporates a 50:50 vol% mixture of toluene and 2-propanol for the
22 removal of the extractable organic carbon. The filter is dried, and the non-extractable organic
23 carbon is removed by thermal desorption under nitrogen at 500 °C. The remaining carbon on the
24 filter, assumed to be elemental carbon, is combusted in an oxidizing atmosphere at 650 °C, and
25 the CO2 produced is detected by coulometry.
26 The other three laboratories using VDI 2465 methods incorporated the VDI 2465/2 method
27 that separates the carbonaceous fractions of the aerosols due to their different thermal stabilities.
28 The sample is first heated in an oxygen free inert gas (either helium or argon) at temperatures of
29 350 and 620 °C over a copper/cerium IV oxide catalyst to remove the organic carbon. The
30 sample is heated at 700 °C in at least 20% oxygen to determine the elemental carbon, and the
31 resulting CO2 is detected by nondispersive infrared spectrophotometry (NDIR). A sixth
April 2002 2-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 laboratory employed a thermal analyzer with a multistep program similar to the VDI2465/2
2 method.
3 Four laboratories utilized different thermal procedures and techniques for detecting the
4 emerging CO2. Of these first nine laboratories, one adapted its technique to correct for
5 pyrolytically formed char. The 10th and 11th laboratories used a thermal optical transmission
6 method (the Sunset Laboratories Inc. instrument) with slightly different temperature programs
7 and atmospheres. The 12th laboratory used a homemade version of the thermal optical
8 transmission method. The 13th laboratory used thermal optical reflectance.
9 The 14th laboratory determined only total carbon using a Shimadzu TOC 5000 with a solid
10 sampler module (SSM-500a). In this method, the sample is combusted at 900 °C over cobalt
11 oxide and platinum catalysts and the evolved CO2 is measured by NDIR. The 15th laboratory
12 analyzed TC by catalytic combustion, using an elemental analyzer CE 440 (Leeman Labs, Inc.)
13 with standard combustion analysis and thermal conductivity detection. Black carbon (BC) was
14 determined by optical transmission!etry, using an aethalometer and an empirical constant of 19
15 cm2 g"1. The 16th laboratory analyzed only BC using an integrating sphere. The 17th laboratory
16 utilized a two-step thermal method, in which the organic material is removed under pure oxygen
17 at 340 °C and the remaining carbon is determined by coulometric titration of the CO2 evolved at
18 1100 °C in a carbon analyzer.
19 Good agreement of the TC results was obtained by all laboratories with only two outliers in
20 the complete data set. The relative standard deviation between laboratories for the TC results
21 were 6.7, 10.6, and 8.8% for the three samples. In contrast, the EC results were much more
22 variable. The relative standard deviation between laboratories for the EC results were 36.6, 24.4,
23 and 45.5% for the three samples. The VDI methods, especially the VDI 2465/2, were found to
24 give generally higher amounts of EC than the thermal-optical methods. This trend was detected
25 for all samples. The authors recognized that uncorrected thermal methods are prone to positive
26 artifacts by charring during pyrolysis. They also noted that when using solvent extraction
27 methods, the dissolution of polymeric aerosol constituents may not be successful. Both of these
28 effects would lead to overestimation of the EC fraction. When the laboratories were grouped
29 according to their methods, the relative standard deviations between laboratories was much
30 smaller. This study demonstrates that the TC measurement can yield similar results from a
31 variety of methods, but the EC measurement is highly dependent upon the method used. The
April 2002 2-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 problems associated with the determination of EC are exacerbated by the lack of a standard
2 reference material.
3 Elemental carbon can also be measured by optical absorption (OA), photoacoustic
4 spectroscopy, and nonextractable mass (Chow et al., 1993). Optical absorption, assumed due
5 entirely to elemental carbon, can be measured by determining light transmission through
6 Teflon-membrane and quartz-fiber filters before and after sampling with a transmission
7 densitometer. Informal intercomparisons among different filter transmission methods have
8 shown high correlations of absorption, but differences of up to a factor of two in absolute values
9 (Watson et al., 1988a,b). These differences are functions of the type of filter, filter loading, the
10 chemical and physical nature of the deposit, the wavelengths of light used, calibration standards,
11 and light diffusing methods. At the current time, there is no agreement on which combination
12 most accurately represents light absorption in the atmosphere. This method is applied with the
13 knowledge that absolute differences in absorption may be found between the measurements made
14 on Teflon-membrane and quartz-fiber filters and with respect to absolute absorption
15 measurements made on the same samples in other laboratories.
16 Black carbon (BC) is also used, in addition to the thermal and thermal/optical methods, for
17 determining EC as a measure of soot (Penner and Novakov, 1996). Both EC and BC define a
18 similar fraction of aerosol; but EC is determined based on thermal properties, whereas BC is
19 based on light-absorption properties. Optical methods for determining BC tend to suffer from
20 calibration problems (Hitzenberger et al., 1996). Lavanchy et al. (1999) compared their EC
21 concentrations as determined from their catalytic thermal oxidation method to BC concentrations
22 determined using an aethalometer operated at the same site; and they found that the instrumental
23 calibration factor provided by the manufacturer was on the order of two times the calibration
24 factor they determined (9.3 ± 0.4 m2g4)- It is possible to calculate a theoretical specific
25 absorption coefficient (Ba) from Mie theory, given a known size distribution and refractive index.
26 The Ba is defined as absorption per mass concentration and can be calculated given the sample
27 filter area, the total deposited mass, and absorption signals for both the loaded and unloaded
28 filters. Often, when no direct measurements are available, values of Ba on the order of 10 n^g"1
29 have been used (Hitzenberger et al., 1996). Typically BC aerosols have values of Ba between
30 3 and 17 mV (Hitzenberger et al., 1996).
April 2002 2-96 DRAFT-DO NOT QUOTE OR CITE
-------
1 Black carbon can be used as an indicator of particles from diesel engines. Therefore, it is
2 important that accurate values for Ba are available. Hitzenberger et al. (1996) investigated the
3 feasibility of using an integrating sphere photometer as an adequate measurement system for the
4 BC content and the absorption coefficient. Based on samples collected during a 10-day period in
5 May 1994, they determined that the usually assumed value of 10 n^g"1 was also applicable to
6 aerosol BC occurring in Vienna.
7 In another study (Hitzenberger et al., 1999), the integrating sphere method was compared to
8 an aethalometer (Hansen et al., 1984), the thermal method of Cachier et al. (1989), and the
9 thermal/optical method of Birch and Gary (1996). The absorption coefficients that were obtained
10 from both the integrating sphere and the aethalometer were comparable. The BC mass
11 concentrations obtained from the aethalometer were 23% of those obtained from the integrating
12 sphere. Compared to the thermal method, the integrating sphere overestimated the BC mass
13 concentrations by 21%. Compared to the thermal/optical method, the integrating sphere was
14 within 5% of the 1:1 line. However, the data were not so well correlated.
15 In 1986, the Carbonaceous Species Methods Comparison Study (CSMCS) was conducted
16 in Los Angeles. The CSMCS was mentioned in the 1996 PM AQCD (U.S. Environmental
17 Protection Agency, 1996a). Hansen and McMurry (1990) compared two very dissimilar methods
18 for aerosol elemental carbon. One involved collection of impactor samples backed by a quartz
19 fiber afterfilter followed by EC analysis by oxidation in helium over a MnO2 catalyst; the other
20 real-time measurements using an aethalometer (an optical absorption technique). They found
21 good agreement between these two very different methods. The CSMCS interlaboratory
22 precision for total carbon was 4.2% (Turpin et al., 2000). However, because the split between
23 OC and EC is operationally defined, there was substantial interlaboratory variability in OC and
24 EC (e.g., 34% for EC [Turpin et al., 1990]). The implications for data analysis are twofold:
25 (1) the analysis method used must be reported with particulate carbon data and (2) comparative
26 analyses should not be conducted with data analyzed by more than one carbon analysis method
27 unless the mutual compatibility of the methods has been demonstrated.
28
29 EC/OC Summary. With the limitations and precautions described above, laboratory
30 analyses for the carbonaceous properties of collected particles have matured to the point where
31 they can be performed with commercially-available instruments following established standard
April 2002 2-97 DRAFT-DO NOT QUOTE OR CITE
-------
1 operating procedures and with traceability to common standards. However, carbon analysis
2 continues to be a subject of research, and carbon methods are currently being compared as part of
3 other studies (e.g., the Atlanta Supersite). The state of the art for soot measurements continues to
4 develop; and, although advances are being made, the definitions of EC and BC continue to be
5 operational and determined by the method employed. Similarly, the distinction between OC and
6 EC is defined operationally. Reports of EC/OC measurements should therefore include mention
7 of the method with which the species were determined. Finally, if possible, all ancillary data
8 should be retained, to allow later comparison to other methods.
9
10 2.2.8 Ionic Species
11 Ion chromatography (1C) is widely used for analyzing ionic species in the water-soluble
12 portion of suspended PM. 1C is the method of choice for the measurement of sulfate, nitrate,
13 ammonium, sodium, and potassium ions for the NAMS program. Aerosol strong acidity, FT, is
14 determined by titration of a water solution of PM collected following a series of annular denuders
15 to remove acid and basic gases with back-up filters to collect NH3 and HNO3 that volatilize from
16 the PM during collection. The 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a)
17 discussed measurement of ions by 1C (Section 4.3.3.1) and of strong acidity (Sections 3.3.1.1 and
18 4.3.3.1); so, no further details will be discussed here.
19
20 2.2.9 Continuous Monitoring
21 The EPA expects that many local environmental agencies will operate continuous PM
22 monitors. All currently available continuous measurements of suspended particle mass share the
23 problem of dealing with semivolatile PM components. So as not to include particle-bound water
24 as part of the mass, the particle-bound water must be removed by heating or dehumidification.
25 However, heating also causes loss of ammonium nitrate and semivolatile organic components.
26 A variety of potential candidates for continuous measurement of mass or chemical components
27 will be discussed in this section.
28
29
30
April 2002 2-98 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.2.9.1 Continuous Measurement of Mass
2 Tapered Element Oscillating Microbalance (TEOM)
3 The advantages of continuous PM monitoring and the designation of the TEOM as an
4 equivalent method for PM10, have led to the deployment of the TEOM at a number of air
5 monitoring sites. The TEOM also is being used to measure PM2 5. The TEOM differs from the
6 federal reference methods for paniculate mass in that it does not require equilibration of the
7 samples at a specified temperature and relative humidity. The TEOM samples at a constant
8 temperature and is typically heated to some temperature higher than the ambient temperature
9 (Meyer et al., 1995; Meyer and Rupprecht, 1996); whereas the FRM samples at the ambient
10 temperature. Thus, the TEOM does not provide data equivalent to the FRM because of losses of
11 volatile species. Volatilization losses in the TEOM sampler can be reduced by operating the
12 instrument at 30 °C, rather than the typical 50 °C specified or by using a Nafion diffusion dryer
13 instead of heating to dehumidify the particles.
14 This difference in operation and implications for fine particle measurements was examined
15 by researchers at CSIRO Atmospheric Research in Australia (Ayers et al., 1999). That group
16 compared 24-h mean PM2 5 mass concentrations as determined by a TEOM and by two manual,
17 gravimetric samplers (a low-volume filter sampler and a MOUDI sampler) in four Australian
18 cities, on 15 days in the winter half-year. The TEOM was operated at 50 °C at one location and
19 at 35 °C at the other three locations. A systematically low TEOM response in comparison to the
20 integrated gravimetric methods was observed. In a comprehensive study, Allen et al. (1997)
21 reported results in which TEOM data collected at 10 urban sites in the United States and Mexico
22 were compared with 24 h integrated mass concentrations for both PM10 and PM2 5. They
23 collected a large data set that included both winter and summer seasons. Allen et al. (1997)
24 concluded that, especially for urban areas, a significant fraction of PM10 could be semivolatile
25 compounds that could be lost from the heated filter in the TEOM leading to a systematic
26 difference between the TEOM and the EPA FRM for PM10. They suggested that this difference
27 is likely to be larger for PM2 5 than it is for PM10 (Allen et al., 1997).
28 In a similar study conducted in Vancouver, British Columbia, the effect of equilibration
29 temperature on PM10 concentrations from the TEOM was examined. Two collocated TEOM
30 monitors operated at 30 and 50 °C, respectively, were operated in the Lower Fraser Valley in
31 British Columbia for a period of approximately 17 months (Mignacca and Stubbs, 1999). A third
April 2002 2-99 DRAFT-DO NOT QUOTE OR CITE
-------
1 TEOM operating at 40 °C was operated for 2 months during this period. They found that, on
2 average, the 1-h average PM10 from the TEOM operating at 30 °C was consistently greater than
3 that from the TEOM operated at 50 °C. For the period during which the third TEOM was
4 operated (at 40 °C), the PM10 from that instrument was between those values for the other two
5 instruments. They also found that the differences in masses were proportional to the PM10
6 loading, and more strongly correlated to the PM10 from the TEOM operated at the lower
7 temperature. They recommended that the TEOM monitors be operated at 40 °C as opposed to
8 operating at 50 °C in summer and 30 °C in winter, in order to avoid introducing a
9 methodological seasonal bias.
10 A new sample equilibration system (SES) was developed to reduce losses of semivolatile
11 species from the PM2 5 TEOM by conditioning the sample stream to lower humidity and
12 temperature (Meyer et al., 2000). The SES utilizes humidity sensors and a Nafion dryer designed
13 for low particle loss. The dryer fits between the flow splitter that follows the size-selective inlet
14 and the sensor unit. A dry purge gas flows over the exterior of the Nafion tubing and allows for
15 self-regeneration. A TEOM with PM2 5 inlet and equipped with an SES was operated at 30 °C
16 alongside another TEOM operating at 50 °C without the SES in Albany, NY, over a 6-day period
17 during a summertime high-temperature, high-relative-humidity episode. The SES maintained the
18 sample air relative humidity under 30%, and the TEOM with the SES generally measured more
19 mass than the other TEOM. The TEOM with SES also was operated alongside an FRM-type
20 sampler for the period of June 6 through September 25, 1999. The correlation between the FRM
21 and TEOM/SES showed a slope of 1.0293 and R2 of 0.9352; whereas the correlation between the
22 FRM and the TEOM without SES and operating at 50 °C showed a slope of 0.8612 and R2 of
23 0.8209. The SES can be installed on existing TEOM monitors.
24 Patashnick et al. (2001) developed a differential TEOM system that is based on a pair of
25 TEOM sensors, each of which is preceded by its own electrostatic precipitator (ESP) and
26 downstream from a common size selective inlet. By alternately switching the ESPs on and off
27 and out of phase with each other, the two sensors measure "effective mass" that includes both the
28 nonvolatile component and the volatile component sampled by the TEOM, less the volatile
29 component that vaporized during the sampling interval. On the sensor side with the ESP turned
30 on, there is no particle collection on that filter so that only volatilization of previously collected
31 particles continues. This allows a correction for the effective mass as measured by the first
April 2002 2-100 DRAFT-DO NOT QUOTE OR CITE
-------
1 sensor by subtracting out the volatilization artifact and leaving the nonvolatile and volatile
2 components of the particulate matter. This system has yet to be well characterized for other
3 biases or interferences such as reactions on the filters, particle collection efficiency of the ESPs,
4 and particle and semivolatile material losses.
5
6 Real-Time Total Ambient Mass Sampler (RAMS)
1 The RAMS, a monitor based on diffusion denuder and TEOM monitor technology, has
8 been developed, validated, and field tested for the real-time determination of total fine PM mass,
9 including semivolatile PM (Eatough et al., 1999a; Obeidi and Eatough, 2002; Obedi et al., 2002;
10 Pang et al., 2001). The RAMS measures the total mass of collected particles including
11 semivolatile species with a TEOM monitor using a "sandwich filter." The sandwich contains a
12 Teflon coated particle collection filter followed by a charcoal-impregnated filter (GIF) to collect
13 any semivolatile species lost from the particles during sampling. Because the instrument
14 measures total mass collected by the sandwich filter, all gas phase compounds that can be
15 adsorbed by a GIF must be removed from the sampling stream prior to the TEOM monitor.
16 Laboratory and field validation data indicate that the precision of fine PM mass determination is
17 better than 10%. The RAMS uses a Nafion dryer to remove particle-bound water from the
18 suspended particles and a particle concentrator to reduce the amount of gas phase organics that
19 must be removed by the denuder. An example of data from the RAMS, the TEOM, and the
20 PC-BOSS is shown in Figure 2-19. This figure also shows the PM2 5 mass from the TEOM as
21 being negative for the hours of 16 to 19. This likely results from the loss of volatile materials
22 from the heated filter.
23
24 Continuous Ambient Mass Monitor (CAMM)
25 Koutrakis and colleagues (Koutrakis et al., 1996; Wang, 1997) have developed CAMM, a
26 technique for the continuous measurement of ambient parti culate matter mass concentration
27 based on the measurement of pressure drop increase with particle loading across a membrane
28 filter. Recently, Sioutas et al. (1999) examined the increase in pressure drop with increasing
29 particle loading on Nuclepore filters. They tested filters with two pore diameters (2 and 5 //m)
30 and filter face velocities ranging from 4 to 52 cm s"1 and examined the effects of relative
31 humidity in the range of 10 to 50%. They found that, for hygroscopic ammonium sulfate
April 2002 2-101 DRAFT-DO NOT QUOTE OR CITE
-------
PC-BOSS (Nonvolatile Material)
PC-BOSS (Lost From Particles)
100
TEOM
at35C
—B— RAMS
at35C
FRM PM25
24 h average
Riverside, CA
D-D.,
D.,
D-r
^ :
-20
21 23 0 1
Time of Day
Figure 2-19. Comparison of mass measurements with collocated RAMS (real-time
data), PC-BOSS (1-h data), FRM PM25 sampler (average of 24-h data),
and a conventional TEOM monitor (real-time data). The semivolatile
fine particulate matter is sampled with the RAMS and PC-BOSS, but
not with the TEOM monitor or the FRM PM2 5 single filter sampler.
The PC-BOSS provides information on both the nonvolatile component
(NV) and the semivolatile organic component (SVOC).
Source: Eatoughetal. (1999a).
1 particles, the change in pressure drop per unit time and concentration was a strong function of
2 relative humidity, decreasing with increasing relative humidity. These results suggest that
3 particulate concentration measurements, similar to the method of Koutrakis et al. (1996) that uses
4 the pressure drop method, may be subject to additional uncertainties if used in an environment
5 where the ambient relative humidity cannot be controlled accurately. The current version of the
6 CAMM (Wang, 1997) uses a particle concentrator, a Nafion dryer, and frequently moves the
April 2002
2-102
DRAFT-DO NOT QUOTE OR CITE
-------
1 filter tape to avoid artifacts due to evaporation of semivolatile components from the active
2 portion of the filter tape which would occur if the atmospheric concentration of the semivolatile
3 components decreased.
4 The CAMMS was recently operated alongside a gravimetric PM method (the Harvard
5 Impactor, or HI) in seven U.S. cities selected for their distinctly different ambient particulate
6 compositions and densities. The correlation between the two methods was high, with an overall
7 r2 of 0.90 and average CAMM/HI ratio of 1.07 (Babich et al., 2000).
8
9 Beta-Gauge Techniques
10 The use of absorption of beta radiation as a indicator of particle mass has been used
11 effectively to measure the mass of equilibrated particulate matter collected on Teflon filters
12 (Jaklevic et al., 1981a; Courtney et al., 1982). The technique also has been used to provide near
13 real-time measurements with time intervals on the order of an hour (Wedding and Weigand,
14 1993). However, real-time beta gauge monitors experience the same problems as other
15 continuous or near real-time particulate matter mass monitoring techniques. Particle-bound
16 water must be removed to reduce the sensitivity of the indicated mass to relative humidity.
17 However, the simplest technique, mild heating, will remove a portion of the ammonium nitrate
18 and the semivolatile organic compounds as well as the particle-bound water.
19 An intercomparison study of two beta gauges at three sites indicated that the Wedding beta
20 gauge and the Sierra Anderson (SA) 1200 PM10 samplers were highly correlated, r > 0.97 (Tsai
21 and Cheng, 1996). The Wedding beta gauge was not sensitive to relative humidity but yielded
22 results approximately 7% lower. This suggests that the mild heating in the beta gauge causes
23 losses comparable to those caused by equilibration, although the differences could result from
24 slight differences in the upper cut points. The Kimoto beta gauge that was operated at ambient
25 temperature was sensitive to relative humidity yielding significantly higher mass concentrations
26 relative to the SA 1200 for RH > 80% than for RH < 80% even though the correlation with the
27 SA 1200 was reasonable (r = 0.94 for RH > 80% and 0.83 for RH < 80%).
28
29 Piezoelectric Microbalance
30 Piezoelectric crystals have mechanical resonances that can be excited by applying an
31 alternating electrical voltage to the crystal. As the resonance frequencies are well defined, such
April 2002 2-103 DRAFT-DO NOT QUOTE OR CITE
-------
1 crystals (quartz in particular) have found applications as secondary time and frequency standards
2 in clocks and watches. As for all mechanical resonators, the resonance frequency is a function of
3 mass. Therefore, by monitoring the resonance frequency in comparison with a second crystal,
4 one can continuously measure the mass deposited on the crystal (Sem et al., 1977; Bowers and
5 Chuan, 1989; Ward and Buttry, 1990; Noel and Topart, 1994). Comparison with a second crystal
6 largely compensates for the effect of temperature changes on the resonance frequency.
7 The piezoelectric principle has been used to measure particle mass by depositing the
8 particles on the crystal surface either by electrostatic precipitation or by impaction (Olin and
9 Sem, 1971). The collection efficiency of either mechanism has to be determined as function of
10 particle size to achieve quantitative measurements. In addition, the mechanical coupling of large
11 particles to the crystal is uncertain. Both single and multi-stage impactors have been used (Olin
12 and Sem, 1971; Fairchild and Wheat, 1984). Quartz crystals have sensitivities of several hundred
13 hertz per microgram. This sensitivity results in the ability to measure the mass concentration of a
14 typical 100 //g/m3 aerosol to within a few percent in less than one minute (Olin and Sem, 1971).
15
16 Coarse Particle Mass
17 The RAMS and CAMM are only appropriate for fine particle measurements (PM2 5 or
18 PMj). However, the TEOM, beta gauge, and piezoelectric microbalance may be used to measure
19 either PM25 or PM10 (or a sample with any specified upper 50% size cut). A pair of such
20 samplers may be used to measure thoracic coarse PM mass concentration (PM10_2 5) by difference
21 between the PM10 and PM2 5 concentrations. However, concerns have been raised concerning the
22 quality of the data from such difference calculations and the resulting potential biases in
23 exposure assessment and risk determinations (Wilson and Suh, 1997; White, 1998). Misra et al.
24 (2001) describe the development and evaluation of a continuous coarse particle monitor (CCPM)
25 that may provide direct measurements of coarse mode PM mass concentrations at short time
26 intervals (on the order of 5-10 min). The basis of the CCPM is enrichment of the coarse particle
27 concentrations through use of virtual impaction while maintaining fine particle concentrations at
28 ambient levels. The resulting aerosol mixture is analyzed using a standard TEOM for which the
29 response is now dominated by the enriched coarse PM mass. The coarse PM concentrations
30 determined from the CCPM were compared to those obtained with a MOUDI, operating with
31 only the 10- and 2.5-micron cutpoint stages, and a Partisol dichotomous sampler. The CCPM
April 2002 2-104 DRAFT-DO NOT QUOTE OR CITE
-------
1 coarse particulate concentrations were highly correlated with both the MOUDI (R2 = 0.88) and
2 the Partisol (R2 = 0.88) coarse PM concentrations. By operating the CCPM at a coarse particle
3 enrichment factor of 25, the coarse PM concentration can be determined a priori without
4 determination of the fine particle concentration, so long as the fine-to-coarse particle
5 concentration ratios are not unusually high (i.e., 4-6). Misra et al. (2001) also concluded from
6 field experiments that the coarse particulate concentrations determined from the CCPM were
7 independent of the ambient fine-to-coarse particulate concentration ratio due to the decrease in
8 particle mass median diameter that should accompany fine-to-coarse particulate concentration
9 ratios during stagnation conditions.
10
11 2.2.9.2 Continuous Measurement of Elemental and Organic Carbon
12 Testing and refinement of models that simulate aerosol concentrations from gas and
13 particle emissions require air quality measurements of approximately 1-h time resolution to
14 reflect the dynamics of atmospheric transport, dispersion, transformation, and removal. Below
15 we describe instruments that have been used to collect and analyze atmospheric organic PM with
16 better than 2-h resolution. These instruments were all present at the Atlanta Supersite
17 experiment during the summer of 1999, and an intercomparison of results is underway.
18 Turpin et al. (1990) describe an in situ, time-resolved analyzer for particulate organic and
19 elemental carbon that can operate on a time cycle as short as 90 min. This analyzer collects
20 particulate matter on a quartz fiber filter mounted in a thermal-optical transmittance carbon
21 analyzer (Turpin et al., 1990). A second quartz fiber filter behind a Teflon filter in a second
22 sampling port may also be analyzed to provide an estimate of the positive sampling artifact (i.e.,
23 gas adsorption on the quartz sampling filter). The organic material in the collected PM is
24 thermally desorbed from the filter at 650 °C and oxidized at 1000 °C over a MnO2 catalyst bed.
25 The evolved CO2 is converted to methane over a nickel catalyst, and the methane is measured in
26 a flame ionization detector. Then the elemental carbon is oxidized at 350 °C in a 98% He-2% O2
27 atmosphere. Correction is made for pyrolytic conversion of some of the organic particulate
28 matter. The instrument was operated with a 2-h resolution during SCAQS in 1987 (Turpin and
29 Huntzicker, 1991;1995), as well as during CSMCS in 1986 (Turpin et al., 1990). By using
30 elemental carbon as a tracer for primary, combustion-generated organic carbon, these authors
31 estimated the contributions of primary sources (i.e., material emitted in particulate form) and
April 2002 2-105 DRAFT-DO NOT QUOTE OR CITE
-------
1 secondary sources (i.e., particulate material formed in the atmosphere) to the total atmospheric
2 particulate organic carbon concentrations in these locations.
3 An automated carbon analyzer with 15-min to 1-h resolution is now commercially available
4 (Rupprecht et al., 1995) and has been operated in several locations including the Atlanta
5 Supersite. It collects samples on a 0. l-//m impactor downstream of an inlet with a 2.5-//m
6 cutpoint. Use of an impactor eliminates gas adsorption that must be addressed when filter
7 collection is used. However, this collection system may experience substantial particle bounce
8 and loss of a sizable fraction of EC since some EC is in particles < 0.2 //m. It is possible that
9 ongoing research, in which particle size is increased by humidification prior to impaction, may
10 result in an improvement in collection efficiency. In the analysis step, carbonaceous compounds
11 are removed by heating in filtered ambient air. Carbonaceous material removed below 340 °C is
12 reported as organic carbon, and material removed between 340 and 750 °C is reported as
13 elemental carbon. Turpin et al. (2000) comment that it would be more appropriate to report
14 carbon values obtained by this method as "low-" and "high-temperature" carbon, because some
15 organics are known to evolve at temperatures greater than 340 °C (e.g., organics from
16 woodsmoke).
17 As discussed earlier, black carbon (BC), a carbon fraction very similar to EC, is most
18 commonly measured using an aethalometer, a commercially available, automated, time-resolved
19 instrument (i.e., 5- to 15-min sample duration) that measures the light attenuation of aerosol
20 particles collected on a filter tape (Hansen et al., 1984). The concentration of elemental carbon is
21 derived from the light absorption measured on a filter using an estimate of the specific absorption
22 (m2/g) of elemental carbon on the filter; the specific absorption value is derived from laboratory
23 and atmospheric tests and is specified by the manufacturer. The specific absorption value could
24 be expected to vary with location, season, and source mix. Comparisons in atmospheric
25 experiments at some locations with EC values measured by thermal methods confirm that the
26 aethalometer provides a statistically meaningful estimate of EC concentration (Allen et al.,
27 1999c; Liousse et al., 1993). For instance, Allen et al. (1999c) found the following statistical
28 relationship for Uniontown, PA, during summer 1990: black carbon (aethaometer) = 0.95*EC
29 (thermal) - 0.2 (r2 = 0.925, n not specified but appears to be >50, EC range from 0 to 9 //g/m3).
30 Another source of error in aethalometer measurements arises from the sampling procedure.
31 Particles are trapped within a three-dimensional filter matrix. Therefore, scattering of
April 2002 2-106 DRAFT-DO NOT QUOTE OR CITE
-------
1 transmitted and reflected light may erroneously be attributed to absorption, thus causing errors in
2 the BC calculation. Ballach et al. (2001) investigated immersing the filter in oil of a similar
3 refractive index as a means to minimize the interferences due to light scattering effects from the
4 filter, similar to a procedure common in microscopy. BC measurements determined using the oil
5 immersion technique were compared to those from an integrating sphere, a polar photometer, and
6 Mie calculations. Aerosols tests included several pure carbon blacks from different generating
7 procedures that were used to calibrate the immersion technique, pure ammonium sulfate aerosol,
8 and external and internal mixtures of ammonium sulfate with varying amounts of carbon blacks.
9 The oil immersion technique was also tested on ambient air samples collected at two different
10 sites in the cities of Frankfurt am Main and Freiburg, Germany. Optical measurements, both of
11 blank and loaded filters, show that the oil immersion technique minimizes scattering losses.
12 Ballach et al. (2001) found that site-related effects were reduced and that there was reasonably
13 good agreement with the other optical techniques as well as with the Mie calculations.
14 Measurement of aerosol light absorption utilizing photoacoustic spectroscopy has been
15 examined as a continuous method for measuring elemental carbon mass concentrations (Petzold
16 and Niessner, 1996; Arnott et al., 1999; 2000). Like the aethalometer, this method measures
17 light absorption; however, unlike most other light absorption methods, the photoacoustic
18 technique does not require a filter. The photoacoustic spectrometer of Arnott and coworkers was
19 demonstrated during the Northern Front Range Air Quality Study and compared to an
20 aethalometer (Moosmuller et al., 1998). Neither the aethalometer nor the photoacoustic
21 spectrometer measure elemental carbon mass directly. Because the photoacoustic spectrometer
22 measures the absorption coefficient directly, the specific absorption efficiency must be known or
23 assumed in order to determine elemental carbon mass. Assuming a light absorption efficiency of
24 10m2 g"1, Arnott et al. (1999) reported a lower detection limit for light absorption of 0.4 M m"1
25 corresponding to a mass concentration of elemental carbon of approximately 40 ng"3.
26
27 2.2.9.3 Continuous Measurements of Nitrate and Sulfate
28 Nitrate
29 An integrated collection and vaporization cell was developed by Stolzenburg and Hering
30 (2000) that provides automated, 10-min resolution monitoring of fine particulate nitrate. In this
31 system, particles are collected by a humidified impaction process and analyzed in place by flash
April 2002 2-107 DRAFT-DO NOT QUOTE OR CITE
-------
1 vaporization and chemiluminescent detection of the evolved nitrogen oxides. In field tests in
2 which the system was collocated with two FRM samplers, the automated nitrate sampler results
3 followed the results from the FRM, but were offset lower. The system also was collocated with a
4 HEADS and a SASS speciation sampler (MetOne Instruments). In all these tests, the automated
5 sampler was well correlated to other samplers with slopes near 1 (ranging from 0.95 for the FRM
6 to 1.06 for the HEADS) and correlation coefficients ranging from 0.94 to 0.996.
7 During the Northern Front Range Air Quality Study in Colorado (Watson et al., 1998), the
8 automated nitrate monitor captured the 12-minute variability in fine particle nitrate
9 concentrations with a precision of approximately ±0.5 //g/m3 (Chow et al., 1998). A comparison
10 with denuded filter measurements followed by ion chromatographic analysis (Chow and Watson,
11 1999) showed agreement within ±0.6 //g/m3 for most of the measurements, but exhibited a
12 discrepancy of a factor of two for the elevated nitrate periods. More recent intercomparisons
13 took place during the 1997 Southern California Ozone Study (SCOS97) in Riverside, CA.
14 Comparisons with 14 days of 24-hour denuder-filter sampling gave a correlation coefficient of
15 R2 = 0.87 and showed no significant bias (i.e., the regression slope is not significantly different
16 from 1). As currently configured, the system has a detection limit of 0.7 //g/m3 and a precision of
17 0.2//g/m3.
18
19 Sulfate
20 Continuous methods for the quantification of aerosol sulfur compounds first remove
21 gaseous sulfur (e.g., SO2, H2S) from the sample stream by a diffusion tube denuder followed by
22 the analysis of particulate sulfur (Cobourn et al., 1978; Durham et al., 1978; Huntzicker et al.,
23 1978; Mueller and Collins, 1980; Tanner et al., 1980). Another approach is to measure total
24 sulfur and gaseous sulfur separately by alternately removing particles from the sample stream.
25 Particulate sulfur is obtained as the difference between the total and gaseous sulfur (Kittelson
26 et al., 1978). The total sulfur content is measured by a flame photometric detector (FPD) by
27 introducing the sampling stream into a fuel-rich hydrogen-air flame (e.g., Stevens et al., 1969;
28 Farwell and Rasmussen, 1976) that reduces sulfur compounds and measures the intensity of the
29 chemiluminescence from electronically excited sulfur molecules (S2*).
30 Because formation of S2* requires two sulfur atoms, the intensity of the chemiluminescence
31 is theoretically proportional to the square of the concentration of molecules that contain a single
April 2002 2-108 DRAFT-DO NOT QUOTE OR CITE
-------
1 sulfur atom. In practice, the exponent is between one and two and depends on the sulfur
2 compound being analyzed (Dagnall et al., 1967; Stevens et al., 1971). Calibrations are
3 performed using both particles and gases as standards. The FPD can also be replaced by a
4 chemiluminescent reaction with ozone that minimizes the potential for interference and provides
5 a faster response time (Benner and Stedman, 1989, 1990).
6 Capabilities added to the basic system include in situ thermal analysis and sulfuric acid
7 speciation (Cobourn et al., 1978; Huntzicker et al., 1978; Tanner et al., 1980; Cobourn and
8 Husar, 1982) ). Sensitivities for particulate sulfur as low as 0.1 //g/m3, with time resolution
9 ranging from 1 to 30 min, have been reported. Continuous measurements of particulate sulfur
10 content have also been obtained by on-line x-ray fluorescence analysis with resolution of 30 min
11 or less (Jaklevic et al., 1981b). During a field-intercomparison study of five different sulfur
12 instruments, Camp et al. (1982) reported four out of five FPD systems agreed to within ±5%
13 during a one-week sampling period.
14
15 2.2.9.4 Continuous Ion Chromatography of Water-Soluble Ions
16 Dasgupta and Slanina have independently developed particle collection systems that grow
17 particles by increasing the relative humidity and collect the particles in an aqueous solution
18 suitable for injection into an ion chromatography (Simon and Dasgupta, 1995; Khlystov et al.,
19 1995). Automation of these systems yield semi-continuous monitors for those ions that can be
20 determined by ion chromatography. A similar system suing a particle size magnifier has been
21 reported by Weber et al. (2001).
22
23 2.2.9.5 Measurements of Individual Particles
24 Recently, several researchers have developed instruments for real-time in situ analysis of
25 single particles (e.g., Noble and Prather, 1996; Gard et al., 1997; Johnson and Wexler, 1995;
26 Silva and Prather, 1997; Thomson and Murphy, 1994). Although the technique varies from one
27 laboratory to another, the underlying principle is to fragment each particle into ions, using either
28 a high-power laser or a heated surface and, then, a time-of-flight mass spectrometer (TOFMS) to
29 measure the ion fragments in a vacuum. Each particle is analyzed in a suspended state in the air
30 stream (i.e., without collection), avoiding sampling artifacts associated with impactors and filters.
31 The technique is called aerosol time-of-flight mass spectrometry (ATOFMS). By measuring both
April 2002 2-109 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
9
10
positive and negative ions from the same particle, information can be obtained about the
chemical composition, not just the elemental composition, of individual particles of known
aerodynamic diameter. This information is especially useful in determining sources of particles.
Noble and Prather (1996) used ATOFMS to provide compositionally resolved particle-size
distributions. Their instrument is capable of analyzing size and chemical composition of 50 to
100 particles/min at typical ambient concentrations and up to 600/min at high particle
concentrations. An example of the type of information that can be determined is
shown in Figure 2-20.
o
O
_0
O
tr
ro
Q.
CD
o>
/////////- Organic
ZZ/////A Marine
Soil
0.2
0.3 0.40.5 0.6 0.70.80.91.0
Aerodynamic Diameter ([jm)
3.0
4.0
Figure 2-20. Size distribution of particles divided by chemical classification into
organic, marine, and crustal.
Source: Noble and Prather (1998).
1 Because particles are analyzed individually, biases in particle sampling (the efficiency of
2 particle transmission into the sensor chamber as a function of size; particle size measurement,
3 and detection of particles prior to fragmentation) represent a major challenge for these
April 2002
2-110
DRAFT-DO NOT QUOTE OR CITE
-------
1 instruments. Moreover, the mass spectrometer has a relatively large variability in ion yields (i.e.,
2 identical samples would yield relatively large differences in mass spectrometer signals [Thomson
3 and Murphy, 1994]); therefore, quantitation is inherently difficult (Murphy and Thomson, 1997).
4 Quantitation will be even more challenging for complex organic mixtures because of the
5 following two reasons: (1) a large number of fragments are generated from each molecule, and
6 (2) ion peaks for organics can be influenced or obscured by inorganic ions (Middlebrook et al.,
7 1998). Nonetheless, scientists have been successful in using these techniques to identify the
8 presence of organics in atmospheric particles and laboratory-generated particles (i.e., as
9 contaminants in laboratory-generated sulfuric acid droplets) as well as the identification of
10 specific compound classes such as PAHs in combustion emissions (Castaldi and Senkan, 1998;
11 Hinz et al., 1994; Middlebrook et al., 1998; Murphy and Thomson, 1997; Neubauer et al., 1998;
12 Noble and Prather, 1998; Reilly et al., 1998; Silva and Prather, 1997). A new multivariate
13 technique for calibration of ATOFMS using microorifice impactors shows promise for
14 simplifying the calibration process (Fergenson et al., 2001). This calibration technique has been
15 applied to gasoline and diesel particles to demonstrate the feasibility of using this technique for
16 the source apportionment of gasoline and diesel particles in an atmospheric mixture (Song et al.,
17 2001).
18 Until recently, ATOFMS systems have only been able to characterize particles that are
19 larger than approximately 0.2 to 0.3 //m in diameter. Wexler and colleagues (Carson et al., 1997;
20 Ge et al., 1998) have developed an ATOFMS instrument that is able to size, count, and provide
21 chemical composition on individual particles ranging in size from 10 nm to 2 //m.
22
23 2.2.9.6 Determination of Aerosol Surface Area in Real Time
24 Aerosol surface area is an important aerosol property for health effects research. However,
25 methods for on-line measurement of surface area are not widely available. Woo et al. (200Ib)
26 used three continuous aerosol sensors to determine aerosol surface area. They used a
27 condensation particle counter (CPC, TSI, Inc., Model 3020), an aerosol mass concentration
28 monitor (MCM, TSI, Inc., Model 8520), and an electrical aerosol detector (BAD, TSI, Inc.,
29 Model 3070) for measuring particle charge concentration. The three sensor signals were inverted
30 to obtain the aerosol size distribution, using a log-normal size distribution model (by minimizing
31 the difference between the measured signals and the theoretical values based upon a size
April 2002 2-111 DRAFT-DO NOT QUOTE OR CITE
-------
1 distribution model, the instrument calibration, and its theoretical responses). The log-normal
2 function was then integrated to calculate the total surface area concentration. Woo et al. (200Ib)
3 demonstrated that this method can give near real-time measurements of aerosol surface area.
4
5 2.2.9.7 Light Scattering
6 A variety of types of nephelometers that integrate aerosol light scattering over various solid
7 angles are available (McMurry, 2000). When used to measure visibility, e.g., to provide pilots
8 with realtime data on visual range, it is desirable to include the light scattering due to particle-
9 bound water. However, when used as an indicator of fine particle mass, it is desirable to exclude
10 particle-bound water. This is frequently done by heating the ambient aerosol to a low reference
11 relative humidity of 40%. However, this heating has the potential of also causing the loss of
12 semivolatile components of the aerosol. The evaporation of ammonium nitrate aerosol in a
13 heated nephelometer has been examined. Bergin et al. (1997) conducted laboratory experiments
14 at low relative humidity (-10%) and as a function of temperature (27-47 °C), mean residence
15 time in the nephelometer, and initial particle size distribution. The evaporation of ammonium
16 nitrate aerosol was also modeled for comparison and was found to describe accurately the
17 decrease in aerosol scattering coefficient as a function of aerosol physical properties and
18 nephelometer operating conditions. Bergin et al. (1997) determined an upper limit estimate of
19 the decrease in the aerosol light scattering coefficient at 450 nm due to evaporation for typical
20 field conditions. The model estimates for their worst-case scenario suggest that the decrease in
21 the aerosol scattering coefficient could be roughly 40%. Under most conditions, however, they
22 estimate that the decrease in aerosol scattering coefficient is generally expected to be less than
23 20%.
24
25 2.2.10 Low Flow Filter Samples for Multiday Collection of Particulate
26 Matter
27 For some purposes, such as demonstrating attainment of an annual standard or as an
28 exposure indicator for epidemiologic studies of chronic health effects, 24-h measurements are not
29 essential. Annual or seasonal averages may be adequate. Multiday sampling techniques can
30 result in lower costs for weighing, chemical analysis, and travel time to change filters. The
31 multiday sampler serves a second purpose. Most commercially available samplers are optimized
April 2002 2-112 DRAFT-DO NOT QUOTE OR CITE
-------
1 for collecting 24-h samples of the PM concentrations found in the U.S., Europe, or Japan. Many
2 cities in other parts of the world have significantly higher PM concentrations. Under these
3 conditions, the 16.7 L/min flow through 37 or 47 mm diameter filters may overload the filter and
4 prevent the sampler from maintaining the prescribed flow rate for 24 h. A low flow sampler with
5 a 0.4 L/min flow rate and a 47 mm diameter filter has been designed by Aerosol Dynamics, Inc.
6 With this sampler, the sample collection time can be chosen to suit the ambient concentration
7 level. This sampler, with a one-week collection period, has been used to characterize PM25 in
8 Beijing, PRC (He et al., 2001). With a two-week collection period, it is being used in a chronic
9 epidemiologic study in southern California, USA (Gauderman, et al., 2000).
10 The sampler, as described by He et al. (2001), has three PM2 5 channels. One channel
11 collects PM on a Teflon filter for gravimetric mass measurement and elemental analysis by XRF.
12 A second channel collects PM on a quartz filter for organic and elemental carbon analysis.
13 A denuder to remove organic gases and a backup filter to collect semivolatile organic compounds
14 may be added. The third channel uses a carbonate denuder to remove acid gases (HNO3 and
15 SO2), a Teflon filter to collect PM for analysis of ions by ion chromatography, and a nylon filter
16 to collect volatilized nitrate. The Teflon filter can also be weighed prior to extraction. Thus, the
17 multiday sampler can provide the information needed for source apportionment by Chemical
18 Element Balance techniques (Watson et al., 1990a,b; U.S. Environmental Protection Agency,
19 2002b).
20 Since PM is commonly sampled on less than daily schedules, the magnitude of sampling
21 errors needs to be considered when quality issues are of concern. For monitoring sites with high
22 day-to-day variability in PM concentrations, an integrated sample may provide a more accurate
23 measurement of the annual average than can be obtained by less-than-everyday sampling
24 schedules. Daily PM data from Spokane, WA were resampled to simulate common sampling
25 schedules, and the sampling error was computed for regulatory and distribution statistics
26 (Rumburg et al., 2001). Probability density functions (pdf s) were fit to the annual daily data to
27 determine the shape of the PM25 concentration distributions. Pdf s were also fit to the less than
28 daily sampling schedules to determine if pdf s could be used to predict the daily high-
29 concentration percentiles. There is an error when using a less than daily sampling schedule for
30 all statistics. The error, expressed as a percentage difference from the everyday sampling, was as
31 large as 1.7, 3.4, and 7.7% for the PM25 mean for l-in-2 day, l-in-3 day, and l-in-6 day
April 2002 2-113 DRAFT-DO NOT QUOTE OR CITE
-------
1 sampling, respectively. For the 98th percentile, the error was as great as 8.8, 18, and 67% for
2 l-in-2 day, l-in-3 day, and l-in-6 day sampling, respectively.
3
4
5 2.3 SUMMARY
6 2.3.1 Atmospheric Physics and Chemistry of Particles
7 Atmospheric particles originate from a variety of sources and possess a range of
8 morphological, chemical, physical, and thermodynamic properties. The composition and
9 behavior of aerosols are linked with those of the surrounding gas. Aerosol may be defined as a
10 suspension of solid or liquid particles in air and includes both the particles and all vapor or gas
11 phase components of air. However, the term aerosol often is used to refer to the suspended
12 particles only.
13 A complete description of the atmospheric aerosol would include an accounting of the
14 chemical composition, morphology, and size of each particle, and the relative abundance of each
15 particle type as a function of particle size. Recent developments in single particle analysis
16 techniques are bringing such a description closer to reality.
17 The diameter of a spherical particle may be determined geometrically, from optical or
18 electron microscopy, by light scattering and Mie theory, or by a particle's behavior (e.g.,
19 electrical mobility or its aerodynamic behavior). However, the various types of diameters may be
20 different, and atmospheric particles often are not spherical. Therefore, particle diameters are
21 described by an "equivalent" diameter. Aerodynamic diameter (i.e., the diameter of a unit
22 density sphere that would have the same terminal settling velocity as the real particle, symbol,
23 Da) is the most widely used equivalent diameter.
24 Atmospheric size distributions show that most atmospheric particles are quite small, below
25 0.1 //m; whereas most of the particle volume (and therefore most of the mass) is found in
26 particles greater than 0.1 //m. An important feature of the mass or volume size distributions of
27 atmospheric particles is their multimodal nature. Volume distributions, measured in ambient air
28 in the United States, are almost always found to be bimodal with a minimum between 1.0 and
29 3.0 //m. The distribution of particles that are mostly larger than the minimum is termed the
30 "coarse" mode. The distribution of particles that are mostly smaller than the minimum is termed
April 2002 2-114 DRAFT-DO NOT QUOTE OR CITE
-------
1 the "fine" mode. Fine-mode particles include both the accumulation mode and the nuclei mode.
2 "Accumulation-mode" particles are that portion of the fine particle fraction with diameters above
3 about 0.1 //m. The nuclei mode, that portion of the fine particle fraction with diameters below
4 about 0.1 //m, can be observed as a separate mode in mass or volume distributions only in clean
5 or remote areas or near sources of new particle formation by nucleation. Toxicologists and
6 epidemiologists use "ultrafme" to refer to particles in the nuclei-mode size range. Aerosol
7 physicists and material scientists tend to use "nanoparticles" to refer to particles in this size
8 range.
9 The aerosol community uses four different approaches or conventions in the classification
10 of particles by size: (1) modes, based on the observed size distributions and formation
11 mechanisms; (2) cut point, usually based on the 50% cut point of the specific sampling device
12 (i.e., the particle size at which 50% of the particles enter and 50% of the particles are rejected);
13 (3) dosimetry or occupational sizes, based on the entrance into various compartments of the
14 respiratory system; and (4) legally specified, regulatory sizes for air quality standards. Over the
15 years, the terms fine and coarse as applied to particle sizes have lost the original precise meaning
16 of fine mode and coarse mode. In any given article, therefore, the meaning of fine and coarse,
17 unless defined, must be inferred from the author's usage. In particular, PM25 and fine-mode
18 particles are not equivalent. In this document, the term "mode" is used with fine and coarse
19 when it is desired to specify the distribution of fine-mode particles or coarse-mode particles as
20 shown in Figures 2-4 and 2-5.
21 Several processes influence the formation and growth of particles. New particles may be
22 formed by nucleation from gas phase material. Particles may grow by condensation as gas phase
23 material condenses onto existing particles. Particles may also grow by coagulation as two
24 particles combine to form one. Gas phase material condenses preferentially on smaller particles,
25 and the rate constant for coagulation of two particles decreases as the particle size increases.
26 Therefore, nuclei mode particles grow into the accumulation mode, but growth of accumulation
27 mode particles into the coarse mode is rare.
28 The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen
29 ions; particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
30 material. Atmospheric PM contains a large number of elements in various compounds and
31 concentrations and hundreds of specific organic compounds. Particulate material can be primary
April 2002 2-115 DRAFT-DO NOT QUOTE OR CITE
-------
1 or secondary. PM is called primary if it is in the same chemical form in which it was emitted
2 into the atmosphere. PM is called secondary if it is formed by chemical reactions in the
3 atmosphere. Primary coarse particles are usually formed by mechanical processes; whereas
4 primary fine particles are emitted from sources either directly as particles or as vapors that
5 rapidly condense to form particles.
6 Most of the sulfate and nitrate and a portion of the organic compounds in atmospheric
7 particles are secondary. Secondary aerosol formation depends on numerous factors including the
8 concentrations of precursors; the concentrations of other gaseous reactive species such as ozone,
9 hydroxyl radical, peroxy radicals, and hydrogen peroxide; atmospheric conditions, including
10 solar radiation and relative humidity; and the interactions of precursors and preexisting particles
11 within cloud or fog droplets or on or in the liquid film on solid particles. As a result, it is
12 considerably more difficult to relate ambient concentrations of secondary species to sources of
13 precursor emissions than it is to identify the sources of primary particles.
14 The lifetimes of particles vary with particle size. Coarse particles can settle rapidly from
15 the atmosphere within minutes or hours and normally travel only short distances. However,
16 when mixed high into the atmosphere, as in dust storms, the smaller-sized, coarse-mode particles
17 may have longer lives and travel greater distances. Accumulation-mode fine particles are kept
18 suspended by normal air motions and have a lower deposition velocity than coarse-mode
19 particles. They can be transported thousands of kilometers and remain in the atmosphere for a
20 number of days. Accumulation-mode particles are removed from the atmosphere primarily by
21 cloud processes. Dry deposition rates are expressed in terms of a deposition velocity that varies
22 with the particle size, reaching a minimum between 0.1 and 1.0 //m aerodynamic diameter.
23 PM is a factor in acid deposition. Particles serve as cloud condensation nuclei and
24 contribute directly to the acidification of rain. In addition, the gas-phase species that lead to dry
25 deposition of acidity are also precursors of particles. Therefore, reductions in SO2 and NOX
26 emissions will decrease both acid deposition and PM concentrations. Sulfuric acid, ammonium
27 nitrate, and organic particles also are deposited on surfaces by dry deposition and can contribute
28 to ecological damage.
29 Particles also reduce visibility and affect radiative balance through scattering and
30 absorption of light. The direct effects of particles in scattering and absorbing light and the
31 indirect effects of particles on clouds impact climate change processes.
April 2002 2-116 DRAFT-DO NOT QUOTE OR CITE
-------
1 2.3.2 Measurement of Atmospheric Particles
2 The decision by the EPA to revise the PM standards by adding daily and yearly standards
3 for PM2 5 has led to a renewed interest in the measurement of atmospheric particles and to a
4 better understanding of the problems in developing precise and accurate measurements of
5 particles. Unfortunately, it is very difficult to measure and characterize particles suspended in
6 the atmosphere.
7 PM monitoring is used to develop information to guide implementation of standards (i.e.,
8 to identify sources of particles; to determine whether or not a standard has been attained; and to
9 determine health, ecological, and radiative effects). Federal Reference Methods (FRM) specify
10 techniques for measuring PM10 and PM25. Particles are collected on filters and mass
11 concentrations are determined gravimetrically. The PM10 FRM sampler consists of a PM10
12 inlet/impactor and a 47-mm Teflon filter with a particle collection efficiency greater than 99.7%.
13 The PM2 5 FRM is similar except that it includes a PM2 5 impactor with an oil-covered impaction
14 substrate to remove particles larger than 2.5 //m. Both techniques provide relatively precise
15 (±10%) methods for determining the mass of material remaining on a Teflon filter after
16 equilibration. Despite considerable progress in measuring the atmospheric PM mass
17 concentration, numerous uncertainties continue to exist as to the relationship between the mass
18 and composition of material remaining on the filter as measured by the FRMs and the mass and
19 composition of material that exists in the atmosphere as suspended PM. There is no reference
20 standard for particles suspended in the atmosphere, nor is there an accepted way to remove
21 particle-bound water without losing some of the semivolatile components of PM such as
22 ammonium nitrate and semivolatile organic compounds. It also is difficult to cleanly separate
23 fine-mode and coarse-mode PM. As a result, EPA defines accuracy for PM measurements in
24 terms of agreement of a candidate sampler with a reference sampler. Therefore,
25 intercomparisons of samplers become very important in determining how well various samplers
26 agree and how various design choices influence what is actually measured.
27 Current filtration-based mass measurements lead to significant evaporative losses of a
28 variety of semivolatile components (i.e., species that exist in the atmosphere in dynamic
29 equilibrium between the condensed phase and gas phase) during and possibly after collection.
30 Important examples include ammonium nitrate, semivolatile organic compounds, and particle-
31 bound water. Loss of these components may significantly affect the quality of the measurement
April 2002 2-117 DRAFT-DO NOT QUOTE OR CITE
-------
1 and can lead to both positive and negative sampling artifacts. Negative artifacts resulting from
2 loss of ammonium nitrate and semivolatile organic compounds may occur during sampling
3 because of changes in temperature, relative humidity, or composition of the aerosol or because of
4 the pressure drop across the filter. Negative artifacts also may occur during handling and storage
5 because of evaporation. Positive artifacts occur when gas-phase compounds (H2O, HNO3, SO2,
6 and organic compounds) absorb onto or react with filter media or collected PM or when some
7 particle-bound water is not removed.
8 Sampling systems for semivolatile PM components make use of denuders to remove the
9 gas-phase fraction and absorptive filters to remove the condensed phase and retain any material
10 that subsequently evaporates from the collected PM. The loss of particulate nitrate may be
11 determined by comparing nitrate collected on a Teflon filter to that collected on a nylon filter
12 (which absorbs nitric acid which evaporates from ammonium nitrate particles) preceded by a
13 denuder to remove gas-phase nitric acid. In two studies in southern California, the PM2 5 mass
14 lost because of volatilization of ammonium nitrate was found to represent 10 to 20% of the total
15 PM25 mass and almost a third of the nitrate. Denuder/absorptive filter sampling systems also
16 have been developed for measuring paniculate phase organic compounds. This technique is an
17 improvement over the filter/adsorbent collection method. However, the denuder systems
18 currently discussed in the literature are not straightforward in their use, and the user must have a
19 thorough understanding of the technology. The FRM for PM2 5 will likely suffer loss of
20 particulate nitrates and semivolatile organic compounds, similar to the losses experienced with
21 other single filter collection systems.
22 It is generally desirable to collect and measure ammonium nitrate and semivolatile organic
23 compounds as part of particulate matter mass. However, it is usually desirable to remove the
24 particle-bound water before determining the mass. In some situations, it may be important to
25 know how much of the suspended particle's mass or volume results from particle-bound water.
26 Calculations and measurements indicate that aerosol water content is strongly dependent on
27 relative humidity and composition. Particle-bound water can represent a significant mass
28 fraction of the PM concentration at relative humidities above 60%. A substantial fraction of
29 accumulation-mode PM is hygroscopic or deliquescent. The more hygroscopic particles tend to
30 contain more sulfates, nitrates, and secondary organic compounds, while the less hygroscopic
31 particles tend to contain more elemental carbon, primary organic compounds, and crustal
April 2002 2-118 DRAFT-DO NOT QUOTE OR CITE
-------
1 components. Fresh, submicron-size soot particles may tend to shrink with increasing relative
2 humidity because of a structural change. The effects of relative humidity on the sorption of
3 semivolatile organic compounds on particles are not well understood. The amount of water
4 sorbed to an atmospheric aerosol may be affected by the presence of an organic film on the
5 particle, which may impede the transport of water across the surface.
6 Fine-mode and coarse-mode particles differ not only in size, but also in formation
7 mechanisms; sources; and chemical, physical, and biological properties. Fine and coarse
8 particles overlap in the intermodal size range (1-2.5 //m Da). As relative humidity increases, fine
9 particles grow into this size range; as relative humidity decreases, more coarse particles may be
10 suspended in this size range. It is desirable to measure fine-mode PM and coarse-mode PM
11 separately in order to properly allocate health effects to either fine-mode PM or coarse-mode PM
12 and to correctly determine sources by factor analysis or chemical mass balance. The selection of
13 a cut point of 2.5 um as a basis for EPA's 1997 NAAQS for fine particles (Federal Register,
14 1997) and its continued use in many health effects studies reflects the importance placed on more
15 complete inclusion of fine-mode particles, while recognizing that intrusion of coarse-mode
16 particles can occur under some conditions with this cut point.
17 In addition to FRM sampling of equilibrated mass to determine compliance with PM
18 standards, EPA requires states to conduct speciation sampling primarily to determine source
19 categories and trends. The current speciation samplers collect PM25 on three filters: (1) a Teflon
20 filter for gravimetric determination of mass and for analysis of heavy elements by X-ray
21 fluorescence; (2) a Nylon filter preceded by a nitric acid denuder for artifact-free determination
22 of nitrate and measurement of other ionic species by ion chromatography; and (3) a quartz filter
23 for measurement of elemental carbon (EC) and organic carbon (OC). In addition, IMPROVE
24 (Interagency Monitoring of Protected Visual Environments) samplers provide information on
25 regional PM background and transport. IMPROVE samplers, in addition to the three types of
26 filters collected by the speciation samplers, also collect a PM10 sample. The IMPROVE and
27 speciation networks use slightly different methods for determination of EC and OC. The two
28 methods agree on total carbon but differ in the split of total carbon into EC and OC. Neither
29 EC/OC method provides for any correction for positive or negative artifacts because of
30 absorption of volatile organic compounds on the quartz filters or evaporation of semivolatile
31 organic compounds from the collected particles.
April 2002 2-119 DRAFT-DO NOT QUOTE OR CITE
-------
1 The EPA expects that monitoring agencies will operate continuous PM monitors. EPA is
2 in the process of providing guidance regarding appropriate continuous monitoring techniques.
3 All currently available techniques for continuous measurements of suspended particle mass such
4 as the integrating nephelometer, the beta-absorption monitor, and the Tapered Element
5 Oscillating Microbalance (TEOM) share the problem of dealing with semivolatile PM
6 components (in order not to include particle-bound water as part of the mass, the particle-bound
7 water must be removed by heating or dehumidification). However, heating also causes
8 ammonium nitrate and semivolatile organic compounds to evaporate. The TEOM monitor
9 operates at a constant, but higher than ambient, temperature to remove particle-bound water.
10 However, the FRM is required to operate at no more than 5 °C above the ambient temperature.
11 Subsequently, much of the particle-bound water is removed during equilibration at 40% relative
12 humidity. This difference in techniques for removal of particle-bound water causes differences
13 in the measured mass concentration between TEOM and FRMs.
14 Several new techniques for continuous PM mass measurements are currently being field
15 tested. The Real-Time Total Ambient Mass Sampler (RAMS) measures the total mass of
16 collected particles including semivolatile species with a TEOM monitor using a "sandwich
17 filter." The sandwich contains a Teflon-coated particle-collection filter followed by a charcoal -
18 impregnated filter to collect any semivolatile species lost from the particles during sampling.
19 The RAMS uses a Nafion dryer to remove particle-bound water from the suspended particles and
20 a particle concentrator to reduce the quantity of gas phase organic compounds that must be
21 removed by the denuder. The Continuous Ambient Mass Monitor (CAMM) estimates ambient
22 particulate matter mass by measurement of the increase in the pressure drop across a membrane
23 filter caused by particle loading. It also uses a Nafion dryer to remove particle-bound water.
24 In addition to continuous mass measurement, a number of techniques for continuous
25 measurement of sulfate, nitrate, or elements are being tested.
26
April 2002 2-120 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Aiginger, H. (1991) Historical development and principles of total reflection X-ray fluorescence analysis (TXRF).
3 Spectrochim. Acta 466:1313-1321.
4 Aiginger, H.; Streli, C. (1997) Total reflection X-ray fluorescence spectrometry (TXRF) for trace element analysis.
5 Spectrosc. Eur. 9: 16-21.
6 Allen, G.; Sioutas, C.; Koutrakis, P.; Reiss, R.; Lurmann, F. W.; Roberts, P. T. (1997) Evaluation of the TEOM®
7 method for measurement of ambient paniculate mass in urban areas. J. Air Waste Manage. Assoc.
8 47:682-689.
9 Allen, J. O.; Paschkewitz, J. S.; Plummer, E. F.; Lafleur, A. L.; Sarofim, A. F.; Smith, K. A. (1999a) Absorption of
10 semi-volatile compounds in oiled impaction substrates: measurement of pyrene absorption. Aerosol Sci.
11 Technol. 30: 16-29.
12 Allen, G. A.; Oh, J. A.; Koutrakis, P. (1999b) Techniques for high quality ambient coarse paniculate mass
13 measurements. J. Air Waste Manage. Assoc. 49: 133-141.
14 Allen, G. A.; Lawrence, J.; Koutrakis, P. (1999c) Field validation of a semi-continuous method for aerosol black
15 carbon (aethalometer) and temporal patterns of summertime hourly black carbon measurements in
16 southwestern PA. Atmos. Environ. 33: 817-823.
17 Allen, R.; Box, M.; Larson, T. V.; Liu, L. J. (2001) A cost-effective weighing chamber for paniculate matter filters.
18 J. Air Waste Manage. Assoc. 51: 1650-1653.
19 American Conference of Governmental Industrial Hygienists. (1994) 1994-1995 threshold limit values for chemical
20 substances and physical agents and biological exposure indices. Cincinnati, OH: American Conference of
21 Governmental Industrial Hygienists.
22 Andrews, E.; Larson, S. M. (1993) Effect of surfactant layers on the size changes of aerosol particles as a function
23 of relative humidity. Environ. Sci. Technol. 27: 857-865.
24 Appel, B. R. (1993) Atmospheric sample analysis and sampling artifacts. In: Willeke, K.; Baron, P. A., eds. Aerosol
25 measurement: principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 233-259.
26 Appel, B. R.; Tokiwa, Y.; Haik, M. (1981) Sampling of nitrates in ambient air. Atmos. Environ. 15: 283-289.
27 Appel, B. R.; Tokiwa, Y.; Haik, M.; Kothny, E. L. (1984) Artifact paniculate sulfate and nitrate formation on filter
28 media. Atmos. Environ. 18: 409-416.
29 Arnott, W. P.; Moosmtiller, H.; Rogers, C. F.; Jin, T.; Bruch, R. (1999) Photoacoustic spectrometer for measuring
30 light absorption by aerosol; instrument description. Atmos. Environ. 33: 2845-2852.
31 Arnott, W. P.; Moosmtiller, H.; Walker, J. W. (2000) Nitrogen dioxide and kerosene-flame soot calibration of
32 photoacoustic instruments for measurement of light absorption by aerosols. Rev. Sci. Instruments
33 71:4545-4552.
34 Ayers, G. P.; Keywood, M. D.; Gras, J. L. (1999) TEOM vs. manual gravimetric methods for determination of
35 PM25 aerosol mass concentrations. Atmos. Environ. 33: 3717-3721.
36 Babich, P.; Davey, M.; Allen, G.; Koutrakis, P. (2000) Method comparisons for paniculate nitrate, elemental
37 carbon, and PM2 5 mass in seven U.S. cities. J. Air Waste Manage. Assoc. 50: 1095-1105.
38 Bacon, J. R.; Chenery, S. R. N.; Ellis, A. T.; Fisher, A.; McMahon, A. W.; Potts, P. J.; Williams, J. G. (1995)
39 Atomic spectrometry update-atomic mass spectrometry and x-ray fluorescence spectrometry. J. Anal. At.
40 Spectrom. 10: 253R-310R.
41 Baldwin, D. P.; Zamzow, D. S.; D'Silva, A. P. (1994) Aerosol mass measurement and solution standard additions
42 for quantitation in laser ablation-inductively coupled plasma atomic emission spectrometry. Anal. Chem.
43 66: 1911-1917.
44 Ballach, J.; Hitzenberger, R.; Schultz, E.; Jaeschke, W. (2001) Development of an improved optical transmission
45 technique for black carbon (BC) analysis. Atmos. Environ. 35: 2089-2100.
46 Baron, P. A.; Willeke, K., eds. (2001) Aerosol measurement: principles, techniques, and applications. New York,
47 NY: John Wiley & Sons.
48 Benner, R. L.; Stedman, D. H. (1989) Universal sulfur detection by chemiluminescence. Anal. Chem.
49 61: 1268-1271.
50 Benner, R. L.; Stedman, D. H. (1990) Field evaluation of the sulfur chemiluminescence detector. Environ. Sci.
51 Technol. 24: 1592-1596.
52 Bergin, M. H.; Ogren, J. A.; Schwartz, S. E.; Mclnnes, L. M. (1997) Evaporation of ammonium nitrate aerosol in a
53 heated nephelometer: implications for field measurements. Environ. Sci. Technol. 31: 2878-2883.
April 2002 2-121 DRAFT-DO NOT QUOTE OR CITE
-------
1 Bertoni, G.; Febo, A.; Perrino, C.; Possanzini, M. (1984) Annular active diffuse sampler: a new device for the
2 collection of organic vapours. Ann. Chim. (Rome) 74: 97-104.
3 Billiet, I; Dams, R.; Hoste, J. (1980) Multielement thin film standards for XRF analysis. X-Ray Spectr. 9: 206-211.
4 Binkowski, F. S.; Ching, J. K. S. (1995) Modeling fine paniculate mass and visibility using the EPA regional
5 paniculate model. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research
6 and Development, National Exposure Research Laboratory; report no. EPA/600/A-95/143. Available from:
7 NTIS, Springfield, VA; PB96-139100.
8 Binkowski, F. S.; Shankar, U. (1995) The regional paniculate matter model. I. Model description and preliminary
9 results. J. Geophys. Res. [Atmos.] 100: 26,191-26,209.
10 Birch, M. E. (1998) Analysis of carbonaceous aerosols: interlaboratory comparison. Analyst 123: 851-857.
11 Birch, M. E.; Gary, R. A. (1996) Elemental carbon-based method for occupational monitoring of paniculate diesel
12 exhaust: methodology and exposure issues. Analyst 121: 1183-1190.
13 Bowers, W. D.; Chuan, R. L. (1989) Surface acoustic-wave piezoelectric crystal aerosol mass monitor. Rev. Sci.
14 Instrum. 60: 1297-1302.
15 Breum, N. O. (2000) The dust holding capacity of porous plastic foam used in particle size-selective sampling.
16 J. Aerosol Sci. 31:379-385.
17 Brook, J. R.; Darin, T. F. (1999) Contribution of nitrate and carbonaceous species to PM25 observed in Canadian
18 cities. J. Air Waste Manage. Assoc. 49: 193-199.
19 Brorstrom, E.; Grennfelt, P.; Lindskog, A. (1983) The effect of nitrogen dioxide and ozone on the decomposition of
20 particle-associated polycyclic aromatic hydrocarbons during sampling from the atmosphere. Atmos. Environ.
21 17:601-605.
22 Bunz, H.; Koyro, M.; Mohler, O.; Saathoff, H. (1996) Temperature dependence of the redistribution mechanisms
23 for volatile components of atmospheric aerosol particles. J. Aerosol Sci. 27(suppl. 1): S653-S654.
24 Cachier, H.; Bremond, M.-P.; Buat-Menard, P. (1989) Determination of atmospheric soot carbon with a simple
25 thermal method. Tellus 41B: 379-390.
26 Cadle, S. H.; Groblicki, P. J. (1982) An evaluation of methods for the determination of organic and elemental
27 carbon in paniculate samples. In: Wolff, G. T.; Klimisch, R. L., eds. Paniculate carbon: atmospheric life
28 cycle. New York, NY: Plenum Press; pp. 89-109.
29 Cadle, S. H.; Mulawa, P.A. (1990) Atmospheric carbonaceous species measurement methods comparison study:
30 General Motors results. Aerosol Sci. Technol. 12: 128-131.
31 Cadle, S. H.; Groblicki, P. J.; Mulawa, P. A. (1983) Problems in the sampling and analysis of carbon particles.
32 Atmos. Environ. 17: 593-600.
33 Cahill, T. A. (1981) Innovative aerosol sampling devices based upon PIXE capabilities. Nucl. Instr. Meth. Phys.
34 Res. 181:473-480.
35 Cahill, T. A. (1985) Proton-induced x-ray emission (PIXE). In: Metals Handbook. Cleveland, OH: American
3 6 Society for Metals.
37 Cahill, T. A.; Eldred, R. A.; Motallebi, N.; Malm, W. C. (1989) Indirect measurement of hydrocarbon aerosols
38 across the United States by nonsulfate hydrogen-remaining gravimetric mass correlations. Aerosol Sci.
39 Technol. 10: 421-429.
40 Cahill, T. A.; Wilkinson, K.; Schnell, R. (1992) Composition analyses of size-resolved aerosol samples taken from
41 aircraft downwind of Kuwait, spring 1991. J. Geophys. Res. [Atmos.] 97: 14,513-14,520.
42 Cahill, T. A.; Wakabayashi, P.; James, T. A. (1996) Chemical states of sulfate at Shenandoah National Park during
43 summer, 1991. Nucl. Instrum. Methods Phys. Res. B 109/110: 542-547.
44 Camp, D. C.; Stevens, R. K.; Cobourn, W. G.; Husar, R. B.; Collins, J. F.; Huntzicker, J. J.; Husar, J. D.; Jaklevic,
45 J. M.; McKenzie, R. L.; Tanner, R. L.; Tesch, J. W. (1982) Intercomparison of concentration results from
46 fine particle sulfur monitors. Atmos. Environ. 16: 911-916.
47 Campbell, A. J.; Humayun, M. (1999) Trace element microanalysis in iron meteorites by laser ablation ICPMS.
48 Anal. Chem. 71:939-946.
49 Carson, P. G.; Johnston, M. V.; Wexler, A. S. (1997) Laser desorption/ionization of ultra fine aerosol particles.
50 Rapid Commun. Mass Spectrom. 11: 993-996.
51 Castaldi, M. J.; Senkan, S. M. (1998) Real-time, ultrasensitive monitoring of air toxics by laser photoionization
52 time-of-flight mass spectroscopy. J. Air Waste Manage. Assoc. 48: 77-81.
53 Chang, M.; Kim, S.; Sioutas, C. (1999) Experimental studies on particle impaction and bounce: effects of substrate
54 design and material. Atmos. Environ. 33: 2313-2322.
April 2002 2-122 DRAFT-DO NOT QUOTE OR CITE
-------
1 Chen, B. T. (1993) Instrument calibration. In: Willeke, K.; Baron, P. A., eds. Aerosol measurement: principles,
2 techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 493-520.
3 Cheng, Y.-S.; Yeh, H.-C. (1979) Particle bounce in cascade impactors. Environ. Sci. Technol. 13: 1392-1396.
4 Chow, J. C. (1995) Measurement methods to determine compliance with ambient air quality standards for
5 suspended particles. J. Air Waste Manage. Assoc. 45: 320-382.
6 Chow, J. C.; Watson, J. G. (1998) Guideline on speciated paniculate monitoring (Draft 3). Research Triangle Park,
7 NC: U.S. Environmental Protection Agency; Office of Air Quality Planning and Standards. Available:
8 http://www.epa.gov/ttn/amtic/files/ambient/pm25/spec/drispec.pdf (20 March 2002).
9 Chow, J. C.; Watson, J. G. (1999) Ion chromatography. In: Landsberger, S.; Creatchman, M., eds. Elemental
10 Analysis of Airborne Particles. Newark, NJ: Gordon and Breach; pp. 97-137.
11 Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993) The DRI
12 thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
13 studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere;
14 April 1991; Vienna, Austria. Atmos. Environ. Part A 27: 1185-1201.
15 Chow, J. C.; Watson, J. G.; Fujita, E. M.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994a) Temporal and spatial
16 variations of PM2.5 and PM10 aerosol in the Southern California Air Quality Study. Atmos. Environ.
17 28:2061-2080.
18 Chow, J. C.; Fujita, E. M.; Watson, J. G.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994b) Evaluation of filter-based
19 aerosol measurements during the 1987 Southern California Air Quality Study. Environ. Monit. Assess.
20 30:49-80.
21 Chow, J. C.; Zielinska, B.; Watson, J. G.; Fujita, E. M.; Richards, H. W.; Neff, W. D.; Dietrich, D.; Hering, S. V
22 (1998) Northern Front Range Air Quality Study. Volume A: Ambient measurements. Fort Collins, CO:
23 Colorado State University, Cooperative Institute for Research in the Atmosphere. Available:
24 http://www.nfraqs.colostate.edu/dri.html [4 February 2002].
25 Chow, J. C.; Watson, J. G.; Hackett, E. I.; Stone, R. H.; Hinsvark, B. A. (2000a) General atomic absorption
26 procedure for trace metals in airborne material collected on filters. In: Lodge, J. P., ed. Methods of Air
27 Sampling and Analysis. 4th ed. Pittsburgh, PA: Air and Waste Management Association.
28 Chow, J. C.; Watson, J. G.; Crow, D.; Lowenthal, D. H.; Merrifield, T. (2000b) Comparison of IMPROVE and
29 NIOSH carbon measurements. Aerosol Sci. Technol. 34: 23-34.
30 Claiborn, C. S.; Finn, D.; Larson, T. V.; Koenig, J. Q. (2000) Windblown dust contributes to high PM2 5
31 concentrations. J. Air Waste Manage. Assoc. 50: 1440-1445.
32 Clark, W. E.; Whitby, K. T. (1975) Measurements of aerosols produced by the photochemical oxidation of SO2 in
33 air. J. Colloid Interface Sci. 51: 477-490.
34 Clarke, A. D. (1992) Atmospheric nuclei in the remote free-troposphere. J. Atmos. Chem. 14: 479-488.
35 Cobourn, W. G.; Husar, R. B. (1982) Diurnal and seasonal patterns of paniculate sulfur and sulfuric acid in
36 St. Louis, July 1977-June 1978. Atmos. Environ. 16: 1441-1450.
37 Cobourn, W. G.; Husar, R. B.; Husar, J. D. (1978) Continuous in situ monitoring of ambient paniculate sulfur using
38 flame photometry and thermal analysis. In: Husar, R. B.; Lodge, J. P., Jr.; Moore, D. J., eds. Sulfur in the
39 atmosphere: proceedings of the international symposium; September 1977; Dubrovnik, Yugoslavia. Atmos.
40 Environ. 12: 89-98.
41 Code of Federal Regulations. (200 la) Appendix J to Part 50~reference method for the determination of paniculate
42 matter as PM10 in the atmosphere. C. F. R. 40: Part 50.
43 Code of Federal Regulations. (200 Ib) Ambient air monitoring reference and equivalent methods. C. F. R.
44 40: Part 53.
45 Code of Federal Regulations. (2001c) Appendix L to Part 50~reference method for the determination of fine
46 paniculate matter as PM2 5 in the atmosphere. C. F. R. 40: Part 50.
47 Code of Federal Regulations. (200 Id) Ambient air quality surveillance for paniculate matter: final rule. C. F. R.
48 40: Part 58.
49 Conner, T.L.; Norris, G.A.; Landis, M.S.; Williams, R.W. (2001) Individual particle analysis of indoor, outdoor,
50 and community samples from the 1998 Baltimore paniculate matter study. Atmos. Environ. 35: 3935-3946.
51 Gotham, W. E.; Bidleman, T. F. (1992) Laboratory investigations of the partitioning of organochlorine compounds
52 between the gas phase and atmospheric aerosols on glass fiber filters. Environ. Sci. Technol. 26: 469-478.
53 Countess, R. J. (1990) Interlaboratory analyses of carbonaceous aerosol samples. Aerosol Sci. Technol.
54 12: 114-121.
April 2002 2-123 DRAFT-DO NOT QUOTE OR CITE
-------
1 Courtney, W. J.; Shaw, R. W.; Dzubay, T. G. (1982) Precision and accuracy of a B gauge for aerosol mass
2 determinations. Environ. Sci. Technol. 16: 236-239.
3 Coutant, R. W.; Callahan, P. J.; Chuang, J. C.; Lewis, R. G. (1992) Efficiency of silicone-grease-coated denuders
4 for collection of polynuclear aromatic hydrocarbons. Atmos. Environ. Part A 26: 2831-2834.
5 Covert, D. S.; Kapustin, V. N.; Quinn, P. K.; Bates, T. S. (1992) New particle formation in the marine boundary
6 layer. J. Geophys. Res. [Atmos.] 97: 20,581-20,589.
7 Cui, W.; Machir, J.; Lewis , L.; Eatough, D. J.; Eatough, N. L. (1997) Fine paniculate organic material at Meadview
8 during the project MOHAVE summer intensive study. J. Air Waste Manage. Assoc. 47: 357-369.
9 Cui, W.; Eatough, D. J.; Eatough, N. L. (1998) Fine paniculate organic material in the Los Angeles Basin -1:
10 assessment of the high-volume Brigham Young University organic sampling system, BIG BOSS. J. Air
11 Waste Manage. Assoc. 48: 1024-1037.
12 Dagnall, R. M.; Thompson, K. C.; West, T. S. (1967) Molecular emission in cool flames I. Analyst 92: 506-512.
13 Dams, R.; Robbins, J. A.; Rahn, K. A.; Winchester, J. W. (1970) Nondestructive neutron activation analysis of air
14 pollution particulates. Anal. Chem. 42: 861-867.
15 Dane, A. D.; Timmermans, P. A. M.; Van Sprang, H. A.; Buydens, L. M. C. (1996) A genetic algorithm for
16 model-free x-ray fluorescence analysis of thin film. Anal. Chem. 68: 2419-2425.
17 Davis, D. W.; Reynolds, R. L.; Tsou, G. C.; Zafonte, L. (1977) Filter attenuation corrections for x-ray fluorescence
18 analysis of atmospheric aerosols. Anal. Chem. 49: 1990-1993.
19 Davis, C. S.; Fellin, P.; Otson, R. (1987) A review of sampling methods for polyaromatic hydrocarbons in air. J. Air
20 Pollut. Control Assoc. 37: 1397-1408.
21 De Boer, D. K. G.; Leenaers, A. J. G.; Van Den Hoogenhof, W. W. (1995) Glancing-incidence x-ray analysis of
22 thin-layered materials: a review. X-ray Spectrom. 24: 91-102.
23 Ding, Y.; Lee, M. L.; Eatough, D. J. (1998a) The determination of total nitrite and N-nitroso compounds in
24 atmospheric samples. Int. J. Environ. Anal. Chem. 69: 243-255.
25 Ding, Y.; Cui, W.; Lee, M. L.; Eatough, D. J. (1998b) Fine paniculate TV-nitroso and nitrite organic compounds in
26 the atmosphere. Appl. Occup. Environ. Hyg. 13: 432-438.
27 Ding, Y.; Pang, Y.; Eatough, D. J.; Eatough, N. L.; Tanner, R. L. (2002) High volume diffusion denuder sampler
28 for the routine monitoring of fine paniculate matter: II. field evaluation of the PC BOSS. Aerosol Sci.
29 Technol.: in press.
30 Dockery, D. W.; Pope, C. A., Ill; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
31 (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 329:
32 1753-1759.
33 Durham, J. L.; Wilson, W. E.; Bailey, E. B. (1978) Application of an SO2-denuder for continuous measurement of
34 sulfur in submicrometric aerosols. Atmos. Environ. 12: 883-886.
35 Dzubay, T. G.; Stevens, R. K. (1975) Ambient air analysis with dichotomous sampler and X-ray fluorescence
36 spectrometer. Environ. Sci. Technol. 9: 663-668.
37 Dzubay, T. G.; Hines, L. E.; Stevens, R. K. (1976) Particle bounce errors in cascade impactors. Atmos. Environ.
38 10:229-234.
3 9 Eatough, D. J. (1999) BOSS, the 5righam Young University Organic Sampling System: determination of paniculate
40 carbonaceous material using diffision denuder sampling technology. In: Lane, D., ed. Gas and particle phase
41 measurements of atmospheric organic compounds. New York, NY: Gordon and Breach Science Publishers.
42 (Advances in environmental, industrial and process control technologies: v. 2).
43 Eatough, D. J.; Wadsworth, A.; Eatough, D. A.; Crawford, J. W.; Hansen, L. D.; Lewis, E. A. (1993)
44 A multiple-system, multi-channel diffusion denuder sampler for the determination of fine-particulate organic
45 material in the atmosphere. Atmos. Environ. Part A 27: 1213-1219.
46 Eatough, D. J.; Tang, H.; Cui, W.; Machir, J. (1995) Determination of the size distribution and chemical
47 composition of fine paniculate semi-volatile organic material in urban environments using diffusion denuder
48 technology. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air pollution
49 and human mortality and morbidity, part II; January 1994; Irvine, CA. Inhalation Toxicol. 7: 691-710.
50 Eatough, D. J.; Eatough, D. A.; Lewis, L.; Lewis, E. A. (1996) Fine paniculate chemical composition and light
51 extinction at Canyonlands National Park using organic paniculate material concentrations obtained with a
52 multisystem, multichannel diffusion denuder sampler. J. Geophys. Res. [Atmos.] 101: 19,515-19,531.
53 Eatough, D. J.; Obeidi, F.; Pang, Y.; Ding, Y.; Eatough, N. L.; Wilson, W. E. (1999a) Integrated and real-time
54 diffusion denuder sampler for PM2 5. Atmos. Environ. 33: 2835-2844.
April 2002 2-124 DRAFT-DO NOT QUOTE OR CITE
-------
1 Eatough, D. I; Pang, Y.; Eatough, N. L. (1999b) Determination of PM2 5 sulfate and nitrate with a PC-BOSS
2 designed for routine sampling for semi-volatile paniculate matter. J. Air Waste Manage. Assoc. 49: 69-75.
3 Fairchild, C. L; Wheat, L. D. (1984) Calibration and evaluation of a real-time cascade impactor. Am. Ind. Hyg.
4 Assoc. 145:205-211.
5 Falconer, R. L.; Bidleman, T. F.; Gotham, W. E. (1995) Preferential sorption of non- and
6 mono-ortho-polychlorinatedbiphenyls to urban aerosols. Environ. Sci. Technol. 29: 1666-1673.
7 Fang, C. P.; McMurry, P. H.; Marple, V. A.; Rubow, K. L. (1991) Effect of flow-induced relative humidity changes
8 on size cuts for sulfuric acid droplets in the microorifice uniform deposit impactor (MOUDI). Aerosol Sci.
9 Technol. 14: 266-277.
10 Farwell, S. O.; Rasmussen, R. A. (1976) Limitations of the FPD and BCD in atmospheric analysis: a review.
11 J. Chromatogr. Sci. 14: 224-234.
12 Fassel, V. A.; Kniseley, R. N. (1974) Inductively coupled plasma-optical emission spectroscopy. Anal. Chem.
13 46: 1110A-1120A.
14 Federal Register. (1987) Revisions to the national ambient air quality standards for paniculate matter. F. R. (July 1)
15 52:24,634-24,669.
16 Federal Register. (1997) National ambient air quality standards for paniculate matter; final rule. F. R. (July 18)
17 62: 38,652-38,752.
18 Federal Register. (1998) National ambient air quality standards for paniculate matter and revised requirements for
19 designation of reference and equivalent methods for PM2 5 and ambient air quality surveillance for paniculate
20 matter; correction. F. R. (February 17) 63: 7710-7718.
21 Fergenson, D. P.; Song, X.-H.; Ramadan, Z.; Allen, J. O.; Hughes, L. S.; Cass, G. R.; Hopke, P. K.; Prather, K. A.
22 (2001) Quantification of ATOFMS data by multivariate methods. Anal. Chem. 73: 3535-3541.
23 Fernandez, F. J. (1989) Atomic absorption spectroscopy. In: Lodge, J. P., Jr., ed. Methods of Air Sampling and
24 Analysis. Chelsea, MI: Lewis Publishers; pp. 83-89.
25 Finn, D.; Rumburg, B.; Claiborn, C.; Bamesberger, L.; Siems, W. F.; Koenig, J.; Larson, T.; Norris, G. (2001)
26 Sampling artifacts from the use of denuder tubes with glycerol based coatings in the measurement of
27 atmospheric paniculate matter. Environ. Sci. Technol. 35: 40-44.
28 Fitz, D. R. (1990) Reduction of the positive organic artifact on quartz filters. Aerosol. Sci. Technol. 12: 142-148.
29 Flocchini, R. G.; Cahill, T. A.; Shadoan, D. J.; Lange, S. J.; Eldred, R. A.; Feeney, P. J.; Wolfe, G. W.; Simmeroth,
30 D. C.; Suder, J. K. (1976) Monitoring California's aerosols by size and elemental composition. Environ. Sci.
31 Technol. 10: 76-82.
32 Friedlander, S. K. (1970) The characterization of aerosols distributed with respect to size and chemical composition.
33 J. Aerosol Sci. 1: 295-307.
34 Friedlander, S. K. (1977) Smoke, dust and haze: fundamentals of aerosol behavior. New York, NY: John Wiley &
35 Sons.
36 Friedlander, S. K.; Yeh, E. K. (1998) The submicron atmospheric aerosol as a carrier of reactive chemical species:
37 case of peroxides. Appl. Occup. Environ. Hyg. 13: 416-420.
3 8 Fuchs, N. A. (1964) The mechanics of aerosols. New York, NY: Pergamon Press.
39 Fuchs, N. A. (1989) The mechanics of aerosols. New York, NY: Dover Publications.
40 Gallorini, M. (1995) Trace element monitoring in atmospheric pollution processes by neutron activation analysis.
41 Microchem. J. 51: 127-137.
42 Card, E.; Mayer, J. E.; Morrical, B. D.; Dienes, T.; Fergenson, D. P.; Prather, K. A. (1997) Real-time analysis of
43 individual atmospheric aerosol particles: design and performance of a portable ATOFMS. Anal. Chem.
44 69:4083-4091.
45 Gauderman, W. J.; McConnell, R.; Gilliland, F.; London, S.; Thomas, D.; Avol, E.; Vora, H.; Berhane, K.;
46 Rappaport, E. B.; Lurmann, F.; Margolis, H. G.; Peters, J. (2000) Association between air pollution and lung
47 function growth in southern California children. Am. J. Respir. Crit. Care Med. 162: 1383-1390.
48 Ge, Z.; Wexler, A. S.; Johnston, M. V. (1998) Laser desorption/ionization of single ultrafine multicomponent
49 aerosols. Environ. Sci. Technol. 32: 3218-3223.
50 Goss, K.-U.; Eisenreich, S. J. (1997) Sorption of volatile organic compounds to particles from a combustion source
51 at different temperatures and relative humidities. Atmos. Environ. 31: 2827-2834.
52 Gray, A. L.; Williams, J. G. (1987a) Oxide and doubly charged ion response of a commercial inductively coupled
53 plasma mass spectrometry instrument. J. Anal. At. Spectrom. 2: 81-82.
April 2002 2-125 DRAFT-DO NOT QUOTE OR CITE
-------
1 Gray, A. L.; Williams, J. G. (1987b) System optimisation and the effect on polyatomic, oxide and doubly charged
2 ion response of a commercial inductively coupled plasma mass spectrometry instrument. J. Anal. At.
3 Spectrom. 2: 599-606.
4 Grosjean, D. (1983) Polycyclic aromatic hydrocarbons in Los Angeles air from samples collected on Teflon, glass
5 and quartz filters. Atmos. Environ. 17: 2565-2573.
6 Grosjean, D.; Parmar, S. S. (1990) Interferences from aldehydes and peroxyacetyl nitrate when sampling urban air
7 organic acids on alkaline traps. Environ. Sci. Technol. 24: 1021-1026.
8 Grosjean, D.; Williams, E. L., II; Grosjean, E.; Novakov, T. (1994) Evolved gas analysis of secondary organic
9 aerosols. Aerosol Sci. Technol. 21: 306-324.
10 Gundel, L. A.; Lane, D. A. (1999) Sorbent-coated diffusion denuders for direct measurement of gas/particle
11 partitioning by semi-volatile organic compounds. In: Lane, D. A., ed. Gas and particle phase measurements
12 of atmospheric organic compounds. Amsterdam, The Netherlands: Gordon and Breach Science Publishers.
13 (Advances in environmental, industrial and process control technologies: v. 2).
14 Gundel, L. A.; Lee, V. C.; Mahanama, K. R. R.; Stevens, R. K.; Daisey, J. M. (1995) Direct determination of the
15 phase distributions of semi-volatile polycyclic aromatic hydrocarbons using annular denuders. Atmos.
16 Environ. 29: 1719-1733.
17 Hanninen, O. O.; Koistinen, K. J. (2002) Quantitative analysis of environmental factors in differential weighing of
18 blank teflon filters. J. Air Waste Manage. Assoc. 52: 134-139.
19 Hansen, A. D. A.; McMurry, P. H. (1990) An intercomparison of measurements of aerosol elemental carbon during
20 the 1986 carbonaceous species method comparison study. J. Air Waste Manage. Assoc. 40: 894-895.
21 Hansen, A. D. A.; Rosen, H.; Novakov, T. (1984) The aethalometer - an instrument for the real-time measurement
22 of optical absorption by aerosol particles. Sci. Total Environ. 36: 191-196.
23 Harman, J. N., Chairman. (1989) Inductively coupled plasma emission spectroscopy. In: Lodge, J. P., Jr., ed.
24 Methods of air sampling and analysis. 3rd ed. Chelsea, MI: Lewis Publishers, Inc.; pp. 89-92.
25 Haupt, O.; Klane, B.; Schaefer, C.; Dannecker, W. (1995) Preparation of quartz fibre filter standards for x-ray
26 fluorescence analysis of aerosol samples. X-Ray Spectrom. 24: 267-275.
27 He, K.; Yang, F.; Ma, Y.; Zhang, Q.; Yao, X.; Chan, C. K.; Cadle, S.; Chan, T.; Mulawa, P. (2001) The
28 characteristics of PM25 in Beijing, China. Atmos. Environ. 35: 4959-4970.
29 Hering, S.; Cass, G. (1999) The magnitude of bias in the measurement of PM25 arising from volatilization of
30 paniculate nitrate from Teflon filters. J. Air Waste Manage. Assoc. 49: 725-733.
31 Hering, S. V.; Appel, B. R.; Cheng, W.; Salaymeh, F.; Cadle, S. H.; Mulawa, P. A.; Cahill, T. A.; Eldred, R. A.;
32 Surovik, M.; Fitz, D.; Howes, J. E.; Knapp, K. T.; Stockburger, L.; Turpin, B. J.; Huntzicker, J. J.; Zhang,
33 X.-Q.; McMurry, P. H. (1990) Comparison of sampling methods for carbonaceous aerosols in ambient air.
34 Aerosol Sci. Technol. 12: 200-213.
3 5 Hinds, W. C. (1999) Aerosol technology: properties, behavior, and measurement of airborne particles. 2nd ed.
3 6 New York, NY: John Wiley & Sons.
37 Hinz, K.-P.; Kaufmann, R.; Spengler, B. (1994) Laser-induced mass analysis of single particles in airborne state.
38 Anal. Chem. 66: 2071-2076.
39 Hitzenberger, R.; Dusek, U.; Berner, A. (1996) Black carbon measurements using an integrating sphere. J. Geophys.
40 Res. [Atmos.] 101: 19,601-19,606.
41 Hitzenberger, R.; Berner, A.; Dusek, U.; Alabashi, R. (1997) Humidity-dependent growth of size-segregated aerosol
42 samples. Aerosol Sci. Technol. 27: 116-130.
43 Hitzenberger, R.; Jennings, S. G.; Larson, S. M.; Dillner, A.; Cachier, H.; Galambos, Z.; Roue, A.; Spain, T. G.
44 (1999) Intercomparison of measurement methods for black carbon aerosols. Atmos. Environ. 33: 2823-2833.
45 Holynska, B.; Ptasinski, J.; Maenhaut, W.; Annegarn, H. J. (1997) Energy-dispersive x-ray fluorescence
46 spectrometer with capillary optics for the chemical analysis of atmospheric aerosols with high time
47 resolution. J. Aerosol Sci. 28: 1455-1464.
48 Houck, J. E.; Chow, J. C.; Watson, J. G.; Simons, C. A.; Pritchett, L. C.; Goulet, J. M.; Frazier, C. A. (1989)
49 Determination of particle size distribution and chemical composition of paniculate matter from selected
50 sources in California: volume I and executive summary (final report). Sacramento, CA: California Air
51 Resources Board.
52 Huntzicker, J. J.; Hoffman, R. S.; Ling, C.-S. (1978) Continuous measurement and speciation of sulfur-containing
53 aerosols by flame photometry. Atmos. Environ. 12: 83-88.
54
April 2002 2-126 DRAFT-DO NOT QUOTE OR CITE
-------
1 Husar, R. B.; Tratt, D. M; Schichtel, B. A.; Falke, S. R.; Li, F.; Jaffe, D.; Gasso, S.; Gill, T.; Laulainen, N. S.;
2 Lu, F.; Reheis, M. C; Chun, Y.; Westphal, D.; Holben, B. N.; Gueymard, C.; McKendiy, I.; Kuring, N.;
3 Feldman, G. C.; McClain, C.; Frouin, R. J.; Merrill, J.; DuBois, D.; Vignola, F.; Murayama, T.; Nickovic, S.;
4 Wilson, W. E.; Sassen, K.; Sugimoto, N.; Malm, W. C. (2001) The Asian dust events of April 1998.
5 J. Geophys. Res. 106: 18,317-18,330.
6 Jackson, K. W.; Mahmood, T. M. (1994) Atomic absorption, atomic emission, and flame emission spectrometry.
7 Anal. Chem. 66: 252R-279R.
8 Jaklevic, J. M.; Loo, B. W.; Goulding, F. S. (1977) Photon-induced X-ray fluorescence analysis using
9 energy-dispersive detector and dichotomous sampler. In: Dzubay, T. G., ed. X-ray fluorescence analysis of
10 environmental samples. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.; pp. 3-18.
11 Jaklevic, J. M.; Gatti, R. C.; Goulding, F. S.; Loo, B. W. (1981a) A p-gauge method applied to aerosol samples.
12 Environ. Sci. Technol. 15: 680-686.
13 Jaklevic, J. M.; Loo, B. W.; Fujita, T. Y. (1981b) Automatic paniculate sulfur measurements with a dichotomous
14 sampler and on-line x-ray fluorescence analysis. Environ. Sci. Technol. 15: 687-690.
15 Jang, M.; Kamens, R. M. (1998) A thermodynamic approach for modeling partitioning of semivolatile organic
16 compounds on atmospheric paniculate matter: humidity effects. Environ. Sci. Technol. 32: 1237-1243.
17 Jang, M.; Kamens, R. M.; Leach, K. B.; Strommen, M. R. (1997) A thermodynamic approach using group
18 contribution methods to model the partitioning of semivolatile organic compounds on atmospheric paniculate
19 matter. Environ. Sci. Technol. 31: 2805-2811.
20 Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. (1996) Particle concentrations, gas-particle partitioning,
21 and species intercorrelations for polycyclic aromatic hydrocarbons (PAH) emitted during biomass burning.
22 Atmos. Environ. 30: 3825-3835.
23 Johansson, T. B.; Askelsson, R.; Johansson, S. A. E. (1970) X-ray analysis: elemental trace analysis at the 10-(12)g
24 level. Nucl. Instrum. Methods 84: 141-143.
25 John, W.; Sethi, V. (1993) Threshold for resuspensionby particle impaction. Aerosol Sci. Technol. 19: 69-79.
26 John, W.; Wall, S. M.; Ondo, J. L. (1988) A new method for nitric acid and nitrate aerosol measurement using the
27 dichotomous sampler. Atmos. Environ. 22: 1627-1635.
28 Johnson, M V.; Wexler, A. S. (1995) MS of individual aerosol particles. Anal. Chem. 67: 721A-726A.
29 Johnson, R. L.; Shah, J. J.; Huntzicker, J. J. (1980) Analysis of organic, elemental, and carbonate carbon in ambient
30 aerosols. In: Verner, S. S. Sampling and analysis of toxic organics in the atmosphere: Proceedings from a
31 symposium; August, 1979; Boulder, CO. Philadelphia, PA: American Society for Testing and Materials,
32 pp. 111-119. [ASTM Special Technical Publication 721].
33 Kamens, R. M.; Coe, D. L. (1997) A large gas-phase stripping device to investigate rates of PAH evaporation from
34 airborne diesel soot particles. Environ. Sci. Technol. 31: 1830-1833.
35 Kamens, R.; Odum, J.; Fan, Z.-H. (1995) Some observations on times to equilibrium for semivolatile polycyclic
36 aromatic hydrocarbons. Environ. Sci. Technol. 29: 43-50.
37 Kamens, R.; Jang, M.; Chien, C.-J.; Leach, K. (1999) Aerosol formation from the reaction of a-pinene and ozone
38 using a gas-phase kinetics-aerosol partitioning model. Environ. Sci. Technol. 33: 1430-1438.
39 Kao, A. S.; Friedlander, S. K. (1995) Frequency distributions of PM10 chemical components and their sources.
40 Environ. Sci. Technol. 29: 19-28.
41 Kavouras, I. G.; Koutrakis, P. (2001) Use of polyurethane foam as the impaction substrate/collection medium in
42 conventional inertia! impactors. Aerosol Sci. Technol. 34: 46-56.
43 Keeler, G. J.; Pirrone, N. (1996) Atmospheric transport and deposition of trace elements to Lake Erie from urban
44 areas. Water Sci. Technol. 33: 259-265.
45 Keeler, G. J.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Pierson, W. R. (1988) Effect of ambient humidity
46 on dichotomous sampler coarse/fine ratios. Atmos. Environ. 22: 1715-1720.
47 Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
48 segregated mass concentration data and ultrafine particle number concentrations in urban areas.
49 Atmos. Environ. 33: 2907-2913.
50 Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (2000) Size distribution and sources of
51 aerosol in Launceston, Australia, during winter 1997. J. Air Waste Manage. Assoc. 50: 418-427.
52 Khlystov, A.; Wyers, G. P.; Slanina, J. (1995) The steam-jet aerosol collector. Atmos. Environ. 29: 2229-2234.
53 Kim, Y. J.; Boatman, J. F.; Gunter, R. L.; Wellman, D. L.; Wilkison, S. W. (1993) Vertical distribution of
54 atmospheric aerosol size distribution over south-central New Mexico. Atmos. Environ. Part A
55 27: 1351-1362.
April 2002 2-127 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kim, B. M; Lester, I; Tisopulos, L.; Zeldin, M. D. (1999) Nitrate artifacts during PM2 5 sampling in the South
2 Coast Air Basin of California. J. Air Waste Manage. Assoc. 49: 142-153.
3 Kim, B. M.; Cassmassi, J.; Hogo, H.; Zeldin, M. D. (2001) Positive organic carbon artifacts on filter medium during
4 PM2 5 sampling in the South Coast Air Basin. Aerosol. Sci. Technol. 34: 35-41.
5 Kirchstetter, T. W.; Corrigan, C. E.; Novakov, T. (2001) Laboratory and field investigation of the adsorption of
6 gaseous organic compounds onto quartz filters. Atmos. Environ. 35: 1663-1671.
7 Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
8 Kittelson, D. B.; McKenzie, R.; Vermeersch, M.; Dorman, F.; Pui, D.; Linne, M.; Liu, B.; Whitby, K. (1978) Total
9 sulfur aerosol concentration with an electrostatically pulsed flame photometric detector system. Atmos.
10 Environ. 12: 105-111.
11 Klouda, G. A.; Klinedinst, D. B.; Steel, E. B.; Benner, B. A., Jr.; Parish, H. J. (1996) Exploring a method to
12 produce an urban dust particle filter standard. J. Aerosol Sci. 27(suppl. 1): S351-S352.
13 Koistinen, K. J.; Kousa, A.; Tenhola, V.; Hanninen, O.; Jantunen, M. J.; Oglesby, L.; Kuenzli, N.; Georgoulis, L.
14 (1999) Fine particle (PM2 5) measurement methodology, quality assurance procedures and pilot results of the
15 EXPOLIS study. J. Air Waste Manage. Assoc. 49: 1212-1220.
16 Koutrakis, P.; Kelly, B. P. (1993) Equilibrium size of atmospheric aerosol sulfates as a function of particle acidity
17 and ambient relative humidity. J. Geophys. Res. [Atmos.] 98: 7141-7147'.
18 Koutrakis, P.; Wolfson, J. M.; Slater, J. L.; Brauer, M.; Spengler, J. D.; Stevens, R. K.; Stone, C. L. (1988)
19 Evaluation of an annular denuder/filter pack system to collect acidic aerosols and gases. Environ. Sci.
20 Technol. 22: 1463-1468.
21 Koutrakis, P.; Wolfson, J. M.; Spengler, J. D.; Stern, B.; Franklin, C. A. (1989) Equilibrium size of atmospheric
22 aerosol sulfates as a function of the relative humidity. J. Geophys. Res. [Atmos.] 94: 6442-6448.
23 Koutrakis, P.; Sioutas, C.; Ferguson, S. T.; Wolfson, J. M.; Mulik, J. D.; Burton, R. M. (1993) Development and
24 evaluation of a glass honeycomb denuder/filter pack system to collect atmospheric gases and particles.
25 Environ. Sci. Technol. 27: 2497-2501.
26 Koutrakis, P.; Wang, P.-Y.; Wolfson, J. M.; Sioutas, C., inventors. (1996) Method and apparatus to measure
27 paniculate matter in gas. U.S. patent 5,571,945. November 5. Available: www.patents.ibm.com [1999,
28 November 24].
29 Krieger, M. S.; Hites, R. A. (1992) Diffusion denuder for the collection of semivolatile organic compounds.
30 Environ. Sci. Technol. 26: 1551-1555.
31 Kusko, B. H.; Cahill, T. A.; Eldred, R. A.; Matsuda, Y.; Miyake, H. (1989) Nondestructive analysis of total
32 nonvolatile carbon by Forward Alpha Scattering Technique (FAST). Aerosol Sci. Technol. 10: 390-396.
33 Lammel, G.; Novakov, T. (1995) Water nucleation properties of carbon black and diesel soot particles.
34 Atmos. Environ. 29: 813-823.
3 5 Landsberger, S. (1988) Improved methodology for the determination of the seven elemental tracer long-distance
36 pollution signatures using thermal and epithermal neutron activation analysis. Anal. Chem. 60: 1842-1845.
37 Landsberger, S.; Wu, D. (1993) Improvement of analytical sensitivities for the determination of antimony, arsenic,
3 8 cadmium, indium, iodine, molybdenum, silicon and uranium in airborne paniculate matter by epithermal
39 neutron activation analysis. J. Radioanal. Nucl. Chem. 167: 219-225.
40 Landsberger, S.; Wu, D.; Vermette, S. J.; Cizek, W. (1997) Intercomparison of IAEA airborne paniculate matter
41 reference material. J. Radioanal. Nucl. Chem. 215: 117-127.
42 Lane, D. A.; Johnson, N. D.; Barton, S. C.; Thomas, G. H. S.; Schroeder, W. H. (1988) Development and
43 evaluation of a novel gas and particle sampler for semivolatile chlorinated organic compounds in ambient air.
44 Environ. Sci. Technol. 22: 941-947.
45 Lavanchy, V. M. H.; Gaggeler, H. W.; Nyeki, S.; Baltensperger, U. (1999) Elemental carbon (EC) and black carbon
46 (BC) measurements with a thermal method and an aethalometer at the high-alpine research station
47 Jungfraujoch. Atmos. Environ. 33: 2759-2769.
48 Lawless, P. A.; Rodes, C. E.; Evans, G.; Sheldon, L.; Creason, J. (2001) Aerosol concentrations during the 1999
49 Fresno Exposure Studies as functions of size, season and meteorology. Aerosol Sci. Technol. 34: 66-74.
50 Lawrence, J. E.; Koutrakis, P. (1994) Measurement of atmospheric formic and acetic acids: methods evaluation and
51 results from field studies. Environ. Sci. Technol. 28: 957-964.
52 Lawrence, J. E.; Koutrakis, P. (1996a) Measurement and speciation of gas and paniculate phase organic acidity in
53 an urban environment: 1. analytical. J. Geophys. Res. [Atmos.] 101: 9159-9169.
54 Lawrence, J.; Koutrakis, P. (1996b) Measurement and speciation of gas and paniculate phase organic acidity in an
55 urban environment: 2. speciation. J. Geophys. Res. [Atmos.] 101: 9171-9184.
April 2002 2-128 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lawson, D. R. (1990) The Southern California Air Quality Study. J. Air Waste Manage. Assoc. 40: 156-165.
2 Lee, W.-M. G.; Shih, P.-M; Wang, C.-S. (1997) The influence of relative humidity on the size of atmospheric
3 aerosol. J. Environ. Sci. Health Part A: Environ. Sci. Eng. Toxic Hazard. Subst. Control 32: 1085-1097.
4 Lewtas, J.; Pang, Y.; Booth, D.; Reimer, S.; Eatough, D. J.; Gundel, L. A. (2001) Comparison of sampling methods
5 for semi-volatile organic carbon associated with PM2 5. Aerosol Sci. Technol. 34: 9-22.
6 Li, W.; Montassier, N.; Hopke, P. K. (1992) A system to measure the hygroscopicity of aerosol particles. Aerosol
7 Sci. Technol. 17: 25-35.
8 Liang, Z.; Chan, C. K. (1997) A fast technique for measuring water activity of atmospheric aerosols. Aerosol Sci.
9 Technol. 26: 255-268.
10 Liang, C.; Pankow, J. F.; Odum, J. R.; Seinfeld, J. H. (1997) Gas/particle partitioning of semivolatile organic
11 compounds to model inorganic, organic, and ambient smog aerosols. Environ. Sci. Technol. 31: 3086-3092.
12 Lin, J. J.; Noll, K. E.; Holsen, T. M. (1994) Dry deposition velocities as a function of particle size in the ambient
13 atmosphere. Aerosol Sci. Technol. 20: 239-252.
14 Lindskog, A.; Brorstrom-Lunden, E.; Alfheim, I; Hagen, I. (1987) Chemical transformation of PAH on airborne
15 particles by exposure to NO2 during sampling: a comparison between two filter media. Sci. Total Environ.
16 61:51-57.
17 Liousse, C.; Cachier, H.; Jennings, S. G. (1993) Optical and thermal measurements of black carbon aerosol content
18 in different environments: variation of the specific attenuation cross-section, sigma (a). Atmos. Environ.
19 27A: 1203-1211.
20 Loo, B. W.; Cork, C. P. (1988) Development of high efficiency virtual impactors. Aerosol Sci. Technol. 9: 167-176.
21 Loo, B. W.; Jaklevic, J. M.; Goulding, F. S. (1976) Dichotomous virtual impactors for large scale monitoring of
22 airborne paniculate matter. In: Liu, B. Y. H., ed. Fine particles: aerosol generation, measurement, sampling,
23 and analysis. New York, NY: Academic Press; pp. 311-350.
24 Lundgren, D. A.; Burton, R. M. (1995) Effect of particle size distribution on the cut point between fine and coarse
25 ambient mass fractions. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air
26 pollution and human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 131-148.
27 Lundgren, D. A.; Hausknecht, B. J.; Burton, R. M. (1984) Large particle size distribution in five U.S. cities and the
28 effect on a new ambient paniculate matter standard (PM10). Aerosol Sci. Technol. 7: 467-473.
29 Lynch, A. J.; McQuaker, N. R.; Brown, D. F. (1980) ICP/AES analysis and the composition of airborne and soil
30 materials in the vicinity of a lead/zinc smelter complex. J. Air Pollut. Control Assoc. 30: 257-260.
31 Mahanama, K. R. R.; Gundel, L. A.; Daisey, J. M. (1994) Selective fluorescence detection of polycyclic aromatic
32 hydrocarbons in environmental tobacco smoke and other airborne particles. Int. J. Environ. Anal. Chem.
33 56:289-309.
34 Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
3 5 concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
36 Mamane, Y.; Willis, R.; Conner, T. (2001) Evaluation of computer-controlled scanning electron microscopy applied
37 to an ambient urban aerosol sample. Aerosol Sci. Technol. 34: 97-107.
38 Marple, V. A.; Chien, C. M. (1980) Virtual impactors: a theoretical study. Environ. Sci. Technol. 14: 976-985.
39 Marple, V. A.; Liu, B. Y. H.; Burton, R. M. (1990) High-volume impactor for sampling fine and coarse particles.
40 J. Air Waste Manage. Assoc. 40: 762-767.
41 Marple, V. A.; Rubow, K. L.; Behm, S. M. (1991) A microorifice uniform deposit impactor (MOUDI): description,
42 calibration, and use. Aerosol Sci. Technol. 14: 434-446.
43 Mclnnes, L. M.; Quinn, P. K.; Covert, D. S.; Anderson, T. L. (1996) Gravimetric analysis, ionic composition, and
44 associated water mass of the marine aerosol. Atmos. Environ. 30: 869-884.
45 McMurry, P. H. (2000) A review of atmospheric aerosol measurements. Atmos. Environ. 34: 1959-1999.
46 McMurry, P. H.; Friedlander, S. K. (1979) New particle formation in the presence of an aerosol. Atmos. Environ.
47 13: 1635-1651.
48 McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
49 aerosols. Atmos. Environ. 23: 497-507.
50 McMurry, P. H.; Woo, K. S.; Weber, R.; Chen, D.-R.; Pui, D. Y. H. (2000) Size distributions of 3-10 nm
51 atmospheric particles: implications for nucleation mechanisms. Philos. Trans. R. Soc. London Ser. A
52 358:2625-2642.
53 McQuaker, N. R.; Kluckner, P. D.; Chang, G. N. (1979) Calibration of an inductively coupled plasma-atomic
54 emission spectrometer for the analysis of environmental materials. Anal. Chem. 51: 888-895.
April 2002 2-129 DRAFT-DO NOT QUOTE OR CITE
-------
1 Meyer, A. S., Jr.; Boyd, C. M. (1959) Determination of water by titration with coulometrically generated Karl
2 Fischer reagent. Anal. Chem. 31: 215-219.
3 Meyer, M. B.; Rupprecht, E. G. (1996) Paniculate matter sampling methods: the importance of standardization.
4 J. Aerosol Sci. 27(suppl. 1): S349-S350.
5 Meyer, M. B.; Rupprecht, E.; Patashnick, H. (1995) Considerations for the sampling and measurement of ambient
6 paniculate matter. In: Paniculate matter: health and regulatory issues: proceedings of an international
7 specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association;
8 pp. 204-210. (A&WMA publication VIP-49).
9 Meyer, M. B.; Patashnick, H.; Ambs, J. L.; Rupprecht, E. (2000) Development of a sample equilibration system for
10 the TEOM continuous PM monitor. J. Air Waste Manage. Assoc. 50: 1345-1349.
11 Middlebrook, A. M.; Murphy, D. M.; Thomson, D. S. (1998) Observations of organic material in individual marine
12 particles at Cape Grim during the first aerosol characterization experiment (ACE 1). J. Geophys. Res.
13 103: 16475-16483.
14 Mignacca, D.; Stubbs, K. (1999) Effects of equilibration temperature on PM10 concentrations from the TEOM
15 method in the Lower Fraser Valley. J. Air Waste Manage. Assoc. 49: 1250-1254.
16 Miller, F. J.; Gardner, D. E.; Graham, J. A.; Lee, R. E., Jr.; Wilson, W. E.; Bachmann, J. D. (1979) Size
17 considerations for establishing a standard for inhalable particles. J. Air Pollut. Control Assoc. 29: 610-615.
18 Misra, C.; Geller, M. D.; Shah, P.; Sioutas, C.; Solomon, P. A. (2001) Development and evaluation of a continuous
19 coarse (PM10-PM25) particle monitor. J. Air Waste Manage. Assoc. 51: 1309-1317.
20 Modey, W. K.; Pang, Y.; Eatough, N. L.; Eatough, D. J. (2001) Fine paniculate (PM2 5) composition in Atlanta,
21 USA: assessment of the particle concentrator-Brigham Young University organic sampling system,
22 PC-BOSS, during the EPA supersite study. Atmos. Environ. 35: 6493-6502.
23 Moosmtiller, Hans; Arnott, W. P.; Rogers, C. F. (1998). Photoacoustic and filter measurements related to aerosol
24 light absorption during the Northern Front Range Air Quality Study (Colorado 1996/1997). Journal of
25 Geophysical Research 103, no. D21: 28,149-28,157.
26 Morawska, L.; Bofinger, N. D.; Kocis, L.; Nwankwoala, A. (1998) Submicrometer and supermicrometer particles
27 from diesel vehicle emissions. Environ. Sci. Technol. 32: 2033-2042.
28 Mueller, P. K.; Collins, J. F. (1980) Development of a paniculate sulfate analyzer. Palo Alto, CA: Electric Power
29 Research Institute; report no. P-1382F.
30 Mueller, P. K.; Fung, K. K.; Heisler, S. L.; Grosjean, D.; Hidy, G. M. (1982) Atmospheric paniculate carbon
31 observations in urban and rural areas of the United States. In: Wolff, G. T.; Klimisch, R. L., eds. Paniculate
32 Carbon: Atmospheric Life Cycle. New York, NY: Plenum Publishing Corporation, 343-370.
33 Mukerjee, S.; Ellenson, W. D.; Lewis, R. G.; Stevens, R. K.; Somerville, M. C.; Shadwick, D. S.; Willis, R. D.
34 (1997) An environmental scoping study in the lower Rio Grande Valley of Texas-Ill, residential
35 microenvironmental monitoring for air, house dust, and soil. Environ. Int. 23: 657-673.
36 Munksgaard, N. C.; Parry, D. L. (1998) Lead isotope ratios determined by ICP-MS: monitoring of mining-derived
37 metal particulates in atmospheric fallout, Northern Territory, Australia. Sci. Total Environ. 217: 113-126.
38 Murphy, D. M.; Thomson, D. S. (1997) Chemical composition of single aerosol particles at Idaho Hill: negative ion
39 measurements. J. Geophys. Res. 102: 6353-6368.
40 Nadkarni, R. A. (1975) Multielement analysis of coal and coal fly ash standards by instrumental neutron activation
41 analysis. Radiochem. Radioanal. Lett. 21: 161-176.
42 Nam, S.; Masamba, W. R. L.; Montaser, A. (1993) Investigation of helium inductively coupled plasma-mass
43 spectrometry for the detection of metals and nonmetals in aqueous solutions. Anal. Chem. 65: 2784-2790.
44 NARSTO. (2002) Fine particle assessment (2nd review draft). Available: http://www.cgenv.com/Narsto/
45 (13 February 2002).
46 National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
47 National Academy Press. 3v.
48 Nelson, M. P.; Zugates, C. T.; Treado, P. J.; Casucio, G. S.; Exline, D. L.; Schlaegle, S. F. (2000) Combining
49 Raman chemical imaging scanning electron mocroscopy (SEM) to characterize ambient fine paniculate.
50 Aerosol Sci. Technol. 34: 108-117.
51 Neubauer, K. R.; Johnson, M. V.; Wexler, A. S. (1998) Humidity effects of the mass spectra of single aerosol
52 particles. Atmos. Environ. 32: 2521-2529.
53
54
April 2002 2-130 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nham, T. T. (1996) Cool plasma technique for the ultra-trace level determination of 56Fe, 40Ca, 39K, 7Li, 24Mg
2 and 23Na for semi-conductor applications. Mulgrave, Victoria, Australia: Varian Australia Pty Ltd; ICP-MS
3 instruments at work no. ICP-MS-7. Available:
4 http://www.varianinc.com/cgi-bin/nav?varinc/docs/osi/icpms/atwork/icpms7&cid=680052 [18 January,
5 2002].
6 Noble, C. A.; Prather, K. A. (1996) Real-time measurement of correlated size and composition profiles of individual
7 atmospheric aerosol particles. Environ. Sci. Technol. 30: 2667-2680.
8 Noble, C. A.; Prather, K. A. (1998) Single particle characterization of albuterol metered dose inhaler aerosol in near
9 real-time. Aerosol Sci. Technol. 29: 294-306.
10 Noble, C. A.; Vanderpool, R. W.; Peters, T. M; McElroy, F. F.; Gemmill, D. B.; Wiener, R. W. (2001) Federal
11 reference and equivalent methods for measuring fine paniculate matter. Aerosol Sci. Technol. 34: 457-464.
12 Noel, M. A.; Topart, P. A. (1994) High-frequency impedance analysis of quartz crystal microbalances. 1. General
13 considerations. Anal. Chem. 66: 484-491.
14 Novakov, T.; Corrigan, C. E.; Penner, J. E.; Chuang, C. C.; Rosario, O.; Mayol Bracero, O. L. (1997) Organic
15 aerosols in the Caribbean trade winds: a natural source? J. Geophys. Res. 102: 21307-21313.
16 Obeidi, F. D.; Eatough, D. J. (2002) Continuous measurement of semi-volatile fine paniculate mass in Provo, UT.
17 Aerosol Sci. Technol. 36: 191-203.
18 Obeidi, F.; Eatough, N. L.; Eatough, D. J. (2002) Use of the RAMS to measure semivolatile fine paniculate matter
19 at Riverside and Bakersfield, California. Aerosol Sci. Technol. 36: 204-216.
20 Ohta, S.; Okita, T. (1990) A chemical characterization of atmospheric aerosol in Sapporo. Atmos. Environ. Part A
21 24A: 815-822.
22 Ohta, S.; Hori, M.; Yamagata, S.; Murao, N. (1998) Chemical characterization of atmospheric fine particles in
23 Sapporo with determination of water content. Atmos. Environ. 32: 1021-1025.
24 Olin, J. G.; Sem, G. J. (1971) Piezoelectric microbalance for monitoring the mass concentration of suspended
25 particles. Atmos. Environ. 5: 653-668.
26 Olmez, I. (1989) Trace element signatures in groundwater pollution. In: Watson, J. G., ed. Receptor models in air
27 resources management: an international specialty conference; February 1988; San Francisco, CA. Pittsburgh,
28 PA: Air & Waste Management Association; pp. 3-11. (APCA transactions series: no. 14).
29 Ondov, J. M. (1996) Paniculate tracers for source attribution: potential for application to California's San Joaquin
30 Valley. J. Aerosol Sci. 27(suppl. 1): S687-S688.
31 Ondov, J. M.; Divita, F., Jr. (1993) Size spectra for trace elements in urban aerosol particles by instrumental neutron
32 activation analysis. J. Radioanal. Nucl. Chem. 167: 247-258.
33 Ondov, J. M.; Kelly, W. R.; Holland, J. Z.; Lin, Z. C.; Wight, S. A. (1992) Tracing fly ash emitted from a coal-fired
34 power plant with enriched rare-earth isotopes: an urban scale test. Atmos. Environ. Part B 26B: 453-462.
35 Ono, D. M.; Hardebeck, E.; Parker, J.; Cox, B. G. (2000) Systematic biases in measured PM10 values with U.S.
36 Environmental Protection Agency-approved samplers at Owens Lake, California. J. Air & Waste Manage.
37 Assoc. 50: 1144-1156.
38 Pak, S. S.; Liu, B. Y. H.; Rubow, K. L. (1992) Effect of coating thickness on particle bounce in inertia! impactors.
39 Aerosol Sci. Technol. 16: 141-150.
40 Pang, Y.; Ren, Y.; Obeidi, F.; Hastings, R.; Eatough, D. J.; Wilson, W. E. (2001) Semi-volatile species inPM25:
41 comparison of integrated and continuous samplers for PM2 s research or monitoring. J. Air Waste Manage.
42 Assoc. 51:25-36.
43 Pang, Y.; Eatough, N. L.; Wilson, J.; Eatough, D. J. (2002a) Effect of semivolatile material on PM2.5 measurement
44 by the PM2.5 federal reference method sampler at Bakersfield, California. Aerosol Sci. Technol.
45 36:289-299.
46 Pang, Y.; Eatough, N. L.; Eatough, D. J. (2002b) PM2.5 semi-volatile organic material at Riverside CA:
47 implications for the PM2.5 federal reference method sampler. Aerosol Sci. Technol. 36: 277-288.
48 Pankow, J. F.; Bidleman, T. F. (1991) Effects of temperature, TSP and per cent non-exchangeable material in
49 determining the gas-particle partitioning of organic compounds. Atmos. Environ. Part A 25: 2241-2249.
50 Pankow, J. F.; Storey, J. M. E.; Yamasaki, H. (1993) Effects of relative humidity on gas/particle partitioning of
51 semivolatile organic compounds to urban paniculate matter. Environ. Sci. Technol. 27: 2220-2226.
52 Parmar, S. S.; Grosjean, D. (1990) Laboratory tests of KI and alkaline annular denuders. Atmos. Environ. Part A
53 24:2695-2698.
54 Patashnick, H.; Rupprecht, G.; Ambs, J. L.; Meyer, M. B. (2001) Development of a reference standard for
55 paniculate matter mass in ambient air. Aerosol Sci. Technol. 34: 42-45.
April 2002 2-131 DRAFT-DO NOT QUOTE OR CITE
-------
1 Pellizzari, E. D.; Krost, K. J. (1984) Chemical transformations during ambient air sampling for organic vapors.
2 Anal. Chem. 56: 1813-1819.
3 Peng, C.; Chan, C.K. (2001) The water cycles of water-soluble organic salts of atmospheric importance.
4 Atmos. Environ. 35: 1183-1192.
5 Penner, J. E.; Novakov, T. (1996) Carbonaceous particles in the atmosphere: a historical perspective to the fifth
6 international conference on carbonaceous particles in the atmosphere. J. Geophys. Res. [Atmos.]
7 101: 19,373-19,378.
8 Perry, K. D.; Cahill, T. A.; Schnell, R. C.; Harris, J. M. (1999) Long-range transport of anthropogenic aerosols to
9 the National Oceanic and Atmospheric Administration baseline station at Mauna Loa Observatory, Hawaii.
10 J. Geophys. Res. Atmos. 104: 18,521-18,533.
11 Peters, A.; Liu, E.; Verrier, R. L.; Schwartz, J.; Gold, D. R.; Mittleman, M.; Baliff, J.; Oh, J. A.; Allen, G.;
12 Monahan, K.; Dockery, D. W. (2000) Air pollution and incidence of cardiac arrhythmia. Epidemiology
13 11: 11-17.
14 Peters, T. M.; Vanderpool, R. W.; Wiener, R. W. (200 la) Design and calibration of the EPA PM2 5 Well Impactor
15 Ninety-Six (WINS), Aerosol Sci. Technol. 34: 389-397.
16 Peters, T. M.; Norris, G. A.; Vanderpool, R. W.; Gemmill, D. B.; Wiener, R. W.; Murdoch, R. W.; McElroy, F. F.;
17 Pitchford, M. (200 Ib) Field performance of PM2 5 federal reference method samplers, Aerosol Sci. Technol.
18 34:433-443.
19 Peters, T. M.; Gussman, R. A.; Kenney, L. C.; Vanderpool, R. W. (2001c) Evaluation of PM25 size selectors used in
20 speciation samplers. Aerosol Sci. Technol. 34: 422-429.
21 Petzold, A., Niessner, R. (1996). Photoacoustic soot sensor for in-situ black carbon monitoring. Applied Physics B,
22 no. 63: 191-197.
23 Petzold, A.; Kopp, C.; Niessner, R. (1997) The dependence of the specific attenuation cross-section on black carbon
24 mass fraction and particle size. Atmos. Environ. 31: 661-672.
25 Pilinis, C.; Seinfeld, J. H.; Grosjean, D. (1989) Water content of atmospheric aerosols. Atmos. Environ.
26 23: 1601-1606.
27 Piorek, S. (1994) Modern, PC based, high resolution portable EDXRF analyzer offers laboratory performance for
28 field, in-situ analysis of environmental contaminants. Nucl. Instr. Meth. Phys. Res. A 353: 528-533.
29 Pirrone, N.; Keeler, G. J. (1996) The Rouge River watershed pollution by trace elements: atmospheric depositions
30 and emission sources. Water Sci. Technol. 33: 267-275.
31 Plantz, M. (1996) Common molecular ion interferences in ICP-MS. Mulgrave, Victoria, Australia: Varian Australia
32 Pty Ltd; ICP-MS instruments at work no. ICP-MS-6. Available:
33 http://www.varianinc.com/cgi-bin/nav?varinc/docs/osi/icpms/atwork/index&cid=680052 [18 January, 2002].
34 Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
3 5 (1995) Paniculate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J.
36 Respir. Crit. Care Med. 151: 669-674.
37 Potts, P. J.; Webb, P. C. (1992) X-ray fluorescence spectrometry. J. Geochem. Explor. 44: 251-296.
3 8 Ranweiler, L. E.; Moyers, J. L. (1974) Atomic absorption procedure for analysis of metals in atmospheric
39 paniculate matter. Environ. Sci. Technol. 8: 152-156.
40 Rao, A. K.; Whitby, K. T. (1977) Nonideal collection characteristics of single stage and cascade impactors.
41 Am. Ind. Hyg. Assoc. J. 38: 174-177.
42 Rao, A. K.; Whitby, K. T. (1978) Non-ideal collection characteristics of inertia! impactors-I. single-stage impactors
43 and solid particles. J. Aerosol Sci. 9: 77-86.
44 Reid, J. S.; Cahill, T. A.; Wakabayashi, P. H.; Dunlap, M. R. (1994) Geometric/aerodynamic equivalent diameter
45 ratios of ash aggregate aerosols collected in burning Kuwaiti well fields. Atmos. Environ. 28: 2227-2234.
46 Reilly, P. T. A.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. (1998) Real-time characterization of the organic
47 composition and size of individual diesel engine smoke particles. Environ. Sci. Technol. 32: 2672-2679.
48 Reist, P. C. (1984) Introduction to aerosol science. New York, NY: Macmillan Publishing Company.
49 Reist, P. C. (1993) Introduction to aerosol science. New York, NY: McGraw-Hill.
50 Risse, U.; Flammenkamp, E.; Kettrup, A. (1996) Determination of aromatic hydrocarbons in air using diffusion
51 denuders. Fresenius J. Anal. Chem. 356: 390-395.
52 Roberts, R. A.; Corkill, J. (1998) Grass seed field smoke and its impact on respiratory health. J. Environ. Health
53 60 (June): 10-16.
54 Rumburg, B.; Alldredge, R.; Claiborn, C. (2001) Statistical distributions of paniculate matter and the error
55 associated with sampling frequency. Atmos. Environ. 35: 2907-2920.
April 2002 2-132 DRAFT-DO NOT QUOTE OR CITE
-------
1 Rupprecht, G.; Patashnick, H.; Beeson, D. E.; Green, R. N.; Meyer, M. B. (1995) A new automated monitor for the
2 measurement of paniculate carbon in the atmosphere. In: Paniculate matter: health and regulatory issues:
3 proceedings of an international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste
4 Management Association; pp. 309-319. (A&WMA publication VIP-49).
5 Sakata, K.; Kawabata, K. (1994) Reduction of fundamental polyatomic ions in inductively coupled plasma mass
6 spectrometry. Spectrochim. Acta49B: 1027-1038.
7 Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
8 literature and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem. 24: 57-109.
9 Saxena, P.; Hildemann, L. M.; McMurry, P. H.; Seinfeld, J. H. (1995) Organics alter hygroscopic behavior of
10 atmospheric particles. J. Geophys. Res. [Atmos.] 100: 18,755-18,770.
11 Schmid, H.; Laskus, L.; Abraham, J. H.; Baltensperger, U.; Lavanchy, V.; Bizjak, M.; Burba, P.; Cachier, H.;
12 Crow, D.; Chow, J.; Gnauk, T.; Even, A.; ten Brink, H. M.; Giesen, K.-P.; Hitzenberger, R.; Hueglin, C.;
13 Maenhaut, W.; Pio, C.; Carvalho, A.; Putaud, J.-Philippe; Toom-Sauntry, D.; Puxbaum, H. (2001) Results of
14 the "carbon conference" international aerosol carbon round robin test stage I. Atmos. Environ.
15 35:2111-2121.
16 Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically with fine particles?
17 J. Air Waste Manage. Assoc. 46: 927-939.
18 Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change.
19 New York, NY: John Wiley & Sons, Inc.
20 Sem, G. J.; Tsurubayashi, K.; Homma, K. (1977) Performance of the piezoelectric microbalance respirable aerosol
21 sensor. Am. Ind. Hyg. Assoc. J. 38: 580-588.
22 Shah, J. J.; Rau, J. A. (1990) Carbonaceous methods comparison study: interlaboratory round robin interpretation
23 results. Sacramento, CA: California Air Resources Board. Available:
24 http://www.arb.ca.gov/research/apr/past/pm.htm (21 March 2002).
25 Silva, P. J.; Prather, K. A. (1997) On-line characterization of individual particles from automobile emissions.
26 Environ. Sci. Technol. 31: 3074-3080.
27 Simon, P. K.; Dasgupta, P. K. (1995) Continuous automated measurement of the soluble fraction of atmospheric
28 paniculate matter. Anal. Chem. 67: 71-78.
29 Sioutas, C.; Koutrakis, P.; Wolfson, J. M. (1994a) Particle losses in glass honeycomb denuder samplers. Aerosol
30 Sci. Technol. 21: 137-148.
31 Sioutas, C.; Koutrakis, P.; Burton, R. M. (1994b) Development of a low cutpoint slit virtual impactor for sampling
32 ambient fine particles. J. Aerosol Sci. 25: 1321-1330.
33 Sioutas, C.; Koutrakis, P.; Burton, R. M. (1994c) A high-volume small cutpoint virtual impactor for separation of
34 atmospheric paniculate from gaseous pollutants. Paniculate Sci. Technol. 12: 207-221.
3 5 Sioutas, C.; Koutrakis, P.; Olson, B. A. (1994d) Development and evaluation of a low cutpoint virtual impactor.
36 Aerosol Sci. Technol. 21:223-235.
37 Sioutas, C.; Koutrakis, P.; Burton, R. M. (1995a) A technique to expose animals to concentrated fine ambient
38 aerosols. Environ. Health Perspect. 103: 172-177.
39 Sioutas, D.; Koutrakis, P.; Ferguson, S. T.; Burton, R. M. (1995b) Development and evaluation of a prototype
40 ambient particle concentrator for inhalation exposure studies. Inhalation Toxicol. 7: 633-644.
41 Sioutas, C.; Koutrakis, P.; Wolfson, J. M.; Azaroff, L. S.; Mulik, J. D. (1996a) Field evaluation of a glass
42 honeycomb denuder/filter pack system to collect atmospheric gases and particles. In: Gay, B. W.; Fuerst,
43 R. G.; Jayanty, R. K. M., eds. Measurement of toxic and related air pollutants: proceedings of a conference;
44 May, 1994; Durham, NC. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
45 Research and Development; report no. EPA/600/R-94/136; pp. 426-435.
46 Sioutas, C.; Wang, P. Y.; Ferguson, S. T.; Koutrakis, P.; Mulik, J. D. (1996b) Laboratory and field evaluation of an
47 improved glass honeycomb denuder/filter pack sampler. Atmos. Environ. 30: 885-895.
48 Sioutas, C.; Koutrakis, P.; Wang, P.-Y.; Babich, P.; Wolfson, J. M. (1999) Experimental investigation of pressure
49 drop with particle loading in Nuclepore filters. Aerosol Sci. Technol. 30: 71-83.
50 Solomon, P. A.; Moyers, J. L.; Fletcher, R. A. (1983) High-volume dichotomous virtual impactor for the
51 fractionation and collection of particles according to aerodynamic size. Aerosol Sci. Technol. 2: 455-464.
52 Solomon, P. A.; Salmon, L. G.; Fall, T.; Cass, G. R. (1992) Spatial and temporal distribution of atmospheric nitric
53 acid and paniculate nitrate concentrations in the Los Angeles area. Environ. Sci. Technol. 26: 1596-1601.
April 2002 2-133 DRAFT-DO NOT QUOTE OR CITE
-------
1 Song, X.-H.; Faber, N. M; Hopke, P. K.; Suess, D. T.; Prather, K. A.; Schauer, J. I; Cass, G. R. (2001) Source
2 apportionment of gasoline and diesel by multivariate calibration based on single particle mass spectral data.
3 Anal. Chim. Acta 446: 329-343.
4 Speer, R. E.; Barnes, H. M.; Brown, R. (1997) An instrument for measuring the liquid water content of aerosols.
5 Aerosol Sci. Technol. 27: 50-61.
6 Stahlschmidt, T.; Schulz, M.; Dannecker, W. (1997) Application of total-reflection X-ray fluorescence for the
7 determination of lead, calcium and zinc in size-fractionated marine aerosols. Spectrochim. Acta B
8 52:995-1001.
9 Stein, S. W.; Turpin, B. J.; Cai, X.; Huang, P.-F.; McMurry, P. H. (1994) Measurements of relative
10 humidity-dependent bounce and density for atmospheric particles using the DMA-impactor technique.
11 Atmos. Environ. 28: 1739-1746.
12 Stevens, R. K.; O'Keeffe, A. E.; Ortman, G. C. (1969) Absolute calibration of a flame photometric detector to
13 volatile sulfur compounds at sub-part-per-million levels. Environ. Sci. Technol. 3: 652-655.
14 Stevens, R. K.; Mulik, J. D.; O'Keefe, A. E.; Krost, K. J. (1971) Gas chromatography of reactive sulfur gases in air
15 at the parts-per-billion level. Anal. Chem. 43: 827-831.
16 Stolzenburg, M. R.; Hering, S. V. (2000) Method for the automated measurement of fine particle nitrate in the
17 atmosphere. Environ. Sci. Technol. 34: 907-914.
18 Streit, N.; Weingartner, E.; Zellweger, C.; Schwikowski, M.; Gaggeler, H. W.; Baltensperger, U. (2000)
19 Characterization of size-fractionated aerosol from the Jungfraujoch (3580 m asl) using total reflection X-ray
20 fluorescence (TXRF). Int. J. Environ. Anal. Chem. 76: 1-16.
21 Strommen, M. R.; Kamens, R. M. (1997) Development and application of a dual-impedance radial diffusion model
22 to simulate the partitioning of semivolatile organic compounds in combustion aerosols. Environ. Sci.
23 Technol. 31:2983-2990.
24 Suarez, A. E.; Caffrey, P. F.; Ondov, J. M.; Thaung, K. C. (1996) Size distribution of Ir-tagged soot from diesel
25 sanitation trucks. J. Aerosol Sci. 27(suppl. 1): S697-S698.
26 Suggs, J. C.; Burton, R. M. (1983) Spatial characteristics of inhalable particles in the Philadelphia metropolitan
27 area. J. AirPollut. Control Assoc. 33: 688-691.
28 Tan, S. H.; Horlick, G. (1986) Background spectral features in inductively coupled plasma mass spectrometry.
29 Appl. Spectres. 40: 445-460.
30 Tang, I. N. (1980) Deliquescence properties and particle size change of hygroscopic aerosols. In: Willeke, K., ed.
31 Generation of aerosols and facilities for exposure experiments. Ann Arbor, MI: Ann Arbor Science
32 Publishers, Inc.; pp. 153-167.
33 Tang, H.; Lewis, E. A.; Eatough, D. J.; Burton, R. M.; Farber, R. J. (1994) Determination of the particle size
34 distribution and chemical composition of semi-volatile organic compounds in atmospheric fine particles with
35 a diffusion denuder sampling system. Atmos. Environ. 28: 939-947.
36 Tanner, R. L.; D'Ottavio, T.; Garber, R.; Newman, L. (1980) Determination of ambient aerosol sulfur using a
37 continuous flame photometric detection system. I. Sampling system for aerosol sulfate and sulfuric acid.
38 Atmos. Environ. 14: 121-127.
39 Thomson, D. S.; Murphy, D. M. (1994) Analyzing single aerosol particles in real time. Chemtech 24: 30-35.
40 Tolocka, M.P.; Peters, T. M.; Vanderpool, R. W.; Chen, F.-L.; Wiener, R. W. (2001a) On the modification of the
41 low flow-rate PM10 dichotomous sampler inlet, Aerosol Sci. Technol. 34: 407-415.
42 Tolocka, M. P.; Solomon, P. A.; Mitchell, W.; Norris, G. A.; Gemmill, D. B.; Wiener, R. W.; Vanderpool, R. W.;
43 Homolya, J. B.; Rice, J. (200Ib) East vs. west in the US: chemical characteristics of PM25 during the winter
44 of 1999. Aerosol Sci. Technol. 34: 88-96.
45 Torok, S. B.; Labar, J.; Schmeling, M.; Van Grieken, R. E. (1998) X-ray spectrometry. Anal. Chem.
46 70:495R-518R.
47 Tsai, C.-J.; Cheng, Y.-H. (1996) Comparison of two ambient beta gauge PM10 samplers. J. Air Waste Manage.
48 Assoc. 46: 142-147.
49 Tsai, C.-J.; Huang, H.-Y. (1995) Atmospheric aerosol sampling by an annular denuder system and a high-volume
50 PM10 sampler. Environ. Int. 21: 283-291.
51 Turner, D. B. (1994) Workbook of atmospheric dispersion estimates - an introduction to dispersion modeling.
52 2nd ed. Boca Raton, FL: Lewis Publishers.
53 Turner, J. R.; Hering, S. V. (1987) Greased and oiled substrates as bounce-free impaction surfaces. J. Aerosol Sci.
54 18:215-224.
April 2002 2-134 DRAFT-DO NOT QUOTE OR CITE
-------
1 Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin:
2 a descriptive analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
3 Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
4 primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
5 Turpin, B. J.; Gary, R. A.; Huntzicker, J. J. (1990) An in situ, time-resolved analyzer for aerosol organic and
6 elemental carbon. Aerosol Sci. Technol. 12: 161-171.
7 Turpin, B. J.; Liu, S.-P.; Podolske, K. S.; Gomes, M. S. P.; Eisenreich, S. J.; McMurry, P. H. (1993) Design and
8 evaluation of a novel diffusion separator for measuring gas/particle distributions of semivolatile organic
9 compounds. Environ. Sci. Technol. 27: 2441-2449.
10 Turpin, B. J.; Huntzicker, J. J.; Hering, S. V. (1994) Investigation of organic aerosol sampling artifacts in the
11 Los Angeles basin. Atmos. Environ. 28: 3061-3071.
12 Turpin, B. J.; Saxena, P.; Andrews, E. (2000) Measuring and simulating paniculate organics in the atmosphere:
13 problems and prospects. Atmos. Environ. 34: 2983-3013.
14 Tyler, G. (1992) ICP vs. AA: how do you choose? Environ. Lab. 31-34.
15 U.S. Court of Appeals, District of Columbia Circuit. (1999) American Trucking Associations, Inc. vs. U.S.
16 Environmental Protection Agency. 175 F3d 1027 (CA DC, 97-1440).
17 U.S. Environmental Protection Agency. (1982) Air quality criteria for paniculate matter and sulfur oxides. Research
18 Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
19 Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
20 U.S. Environmental Protection Agency. (1996a) Air quality criteria for paniculate matter. Research Triangle Park,
21 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
22 U.S. Environmental Protection Agency. (1996b) Review of the national ambient air quality standards for paniculate
23 matter: policy assessment of scientific and technical information. OAQPS staff paper. Research Triangle
24 Park, NC: Office of Air Quality Planning and Standards; report no. EPA/452/R-96-013. Available from:
25 NTIS, Springfield, VA; PB97-115406REB.
26 U.S. Environmental Protection Agency (1998) Monitoring PM2 5 in ambient air using designated reference of class I
27 equivalent methods. In Quality Assurance Handbook for Air Pollution Measurement Systems, Volume II,
28 Part II, Section 2.12.
29 U.S. Environmental Protection Agency. (1999) Paniculate matter (PM2 5) speciation guidance. Final draft. Edition 1.
30 Research Triangle Park, NC: Office of Air Quality Planning and Standards. Available:
31 www.epa.gov/ttn/amtic/files/ambient/pm25/spec/specfinl.pdf [2000, October 11].
32 U.S. Environmental Protection Agency. (2001) List of designated reference and equivalent methods. Available:
33 http://www.epa.gov/ttn/amtic/files/ambient/criteria/repmlist.pdf [05 March 02].
34 U.S. Environmental Protection Agency. (2002a) Ambient monitoring technology information center (AMTIC).
3 5 Washington, DC: Office of Air Quality Planning and Standards; Monitoring and Quality Assurance Group.
36 Available: http://www.epa.gov/ttn/amtic/ (20 March 2002).
37 U.S. Environmental Protection Agency. (2002b) Support center for regulatory air models. Washington, DC: Office
38 of Air Quality Planning and Standards. Available: http://www.epa.gov/scram001 (21 March 2002).
39 Van Loy, M.; Bahadori, T.; Wyzga, R.; Hartsell, B.; Edgerton, E. (2000) The Aerosol Research and Inhalation
40 Epidemiology Study (ARIES): PM2 5 mass and aerosol component concentrations and sampler
41 intercomparisons. J. Air Waste Manage. Assoc. 50: 1446-1458.
42 Vanderpool, R. W.; Peters, T. M.; Natarajan, S.; Gemmill, D. B.; Wiener, R. W. (2001) Evaluation of the loading
43 characteristics of the EPA WINS PM25 separator. Aerosol Sci. Technol. 34: 444-456.
44 Vasilou, J. G.; Sorenson, D.; McMurry, P. H. (1999) Sampling at controlled relative humidity with a cascade
45 impactor. Atmos. Environ. 30: 1049-1056.
46 Venkataraman, C.; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons
47 and elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ. Sci.
48 Technol. 28: 555-562.
49 Vincent, J. H. (1989) Aerosol sampling: science and practice. New York, NY: Wiley.
50 Vincent, J. H. (1995) Aerosol science for industrial hygienists. Oxford, United Kingdom: Pergamon.
51 Wang, P. Y. (1997) Continuous aerosol mass measurement by flow obstruction. Boston, MA: Harvard University,
52 School of Public Health.
53 Wang, H.-C.; John, W. (1987) Comparative bounce properties of particle materials. Aerosol Sci. Technol.
54 7:285-299.
April 2002 2-135 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wang, H.-C.; John, W. (1988) Characteristics of the Berner impactorfor sampling inorganic ions. Aerosol Sci.
2 Technol. 8: 157-172.
3 Ward, M. D.; Buttry, D. A. (1990) In situ interfacial mass detection with piezoelectric transducers. Science
4 249: 1000-1007.
5 Watson, J. G.; Chow, J. C; Richards, L. W.; Neff, W. D.; Andersen, S. R.; Dietrich, D. L.; Houck, J. E.; Olmez, I.
6 (1988a) The 1987-88 metro Denver brown cloud study volume II: measurements. Reno, NV: Desert
7 Research Institute; final report no. 8810.1F2 (2 of 3).
8 Watson, J. G.; Chow, J. C.; Richards, L. W.; Neff, W. D.; Andersen, S. R.; Dietrich, D. L.; Houck, J. E.; Olmez, I.
9 (1988b) The 1987-88 metro Denver brown cloud study volume III: data interpretation. Reno, NV: Desert
10 Research Institute; final report no. 8810.1F3 (3 of 3).
11 Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
12 (1990a) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
13 Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B.; Nguyen, Q. T.; Meyer, E. L.; Pace, T. G.
14 (1990b) Receptor model technical series, v. Ill (1989 revision): CMB7 user's manual. Research Triangle
15 Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; report no.
16 EPA-450/4-90-004. Available: http://www.epa.gov/scram001/ [19 February 2002].
17 Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M.; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994)
18 Chemical mass balance source apportionment of PM10 during the Southern California Air Quality Study.
19 Aerosol Sci. Technol. 21: 1-36.
20 Watson, J. G.; Fujita, E. M.; Chow, J. C.; Zielinska, B.; Richards, L. W.; Neff, W.; Dietrich, D. (1998) Northern
21 front range air quality study. Final report. Fort Collins, CO: Colorado State University, Cooperative
22 Institute for Research in the Atmosphere. Available: http://www.nfraqs.colostate.edu/index2.html (16 Jan
23 2002).
24 Watson, J. G.; Chow, J. C.; Frazier, C. A. (1999) X-ray fluorescence analysis of ambient air samples. In:
25 Landsberger, S.; Creatchman, M., eds. Elemental Analysis of Airborne Particles. Newark, NJ: Gordon and
26 Breach; pp. 67-96.
27 Weber, R. J.; Orsini, D.; Daun, Y.; Lee, Y.-N.; kotz, P. J.; Brechtel, F. (2001) A particle-into-liquid collector for
28 rapid measurement of aerosol bulk chemical composition. Aerosol Sci. Technol. 35: 718-727.
29 Wedding, J. B.; Weigand, M. A. (1993) An automatic particle sampler with beta gauging. J. Air Waste Manage.
30 Assoc. 43:475-479.
31 Weingartner, E.; Burtscher, H.; Baltensperger, U. (1997) Hygroscopic properties of carbon and diesel soot particles.
32 Atmos. Environ. 31: 2311-2327.
33 Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles.
34 Atmos. Environ. 33: 2301-2312.
35 Wesolowski, J. J.; John, W.; Devor, W.; Cahill, T. A.; Feeney, P. J.; Wolfe, G.; Flocchini, R. (1977) Collection
36 surfaces of cascade impactors. In: Dzubay, T. G., ed. X-ray fluorescence analysis of environmental
37 samples. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.; pp. 121-131.
38 Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
39 Whitby, K. T.; Sverdrup, G. M. (1980) California aerosols: their physical and chemical characteristics. In: Hidy,
40 G. M.; Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. J., eds. The character and origins of
41 smog aerosols: a digest of results from the California Aerosol Characterization Experiment (ACHEX).
42 New York, NY: John Wiley & Sons, Inc.; pp. 477-517. (Advances in environmental science and
43 technology: v. 9).
44 Whitby, K. T.; Charlson, R. E.; Wilson, W. E.; Stevens, R. K. (1974) The size of suspended particle matter in air.
45 Science (Washington, DC) 183: 1098-1099.
46 White, W. H. (1998) Statistical considerations in the interpretation of size-resolved paniculate mass data. J. Air &
47 Waste Manage. Assoc. 48: 454-458.
48 Willeke, K.; Baron, P. A., eds. (1993) Aerosol measurement: principles, techniques and applications. New York,
49 NY: Van Nostrand Reinhold Publishers.
50 Willeke, K.; Whitby, K. T. (1975) Atmosperic aerosols: size distribution interpretation. J. AirPollut. Control Assoc.
51 25:529-534.
52 Williams, E. L., II; Grosjean, D. (1990) Removal of atmospheric oxidants with annular denuders. Environ. Sci.
53 Technol. 24:811-814.
54
55
April 2002 2-136 DRAFT-DO NOT QUOTE OR CITE
-------
1 Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
2 (2000) The 1998 Baltimore Paniculate Matter Epidemiology-Exposure Study: part 1. Comparison of
3 ambient, residential outdoor, indoor and apartment paniculate matter monitoring. J. Exposure Anal.
4 Environ. Epidemiol. 10: 518-532.
5 Wilson, W. E. (1995) Aerosol exposure, physics, and chemistry. Inhalation Toxicol. 7: 769-772.
6 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
7 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
8 Wilson, W. E.; Spiller, L. L.; Ellestad, T. G.; Lamothe, P. J.; Dzubay, T. G.; Stevens, R. K.; Macias, E. S.; Fletcher,
9 R. A.; Husar, J. D.; Husar, R. B.; Whitby, K. T.; Kittelson, D. B.; Cantrell, B. K. (1977) General Motors
10 sulfate dispersion experiment: summary of EPA measurements. J. AirPollut. Control Assoc. 27: 46-51.
11 Woo, K. S.; Chen, D. R.; Pui, D. Y. H.; McMurry, P. H. (2001a) Measurement of Atlanta aerosol size distributions:
12 observations of ultrafine particle events. Aerosol Sci. Technol. 34: 75-87.
13 Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; Wilson, W. E. (2001b) Use of continuous measurements of integral aerosol
14 parameters to estimate particle surface area. Aerosol Sci. Technol. 34: 57-65.
15 Wu, C. C.; Suarez, A. E.; Lin, Z.; Kidwell, C. B.; Borgoul, P. V.; Caffrey, P. F.; Ondov, J. M.; Sattler, B. (1998)
16 Application of an Ir tracer to determine soot exposure to students commuting to school on Baltimore public
17 buses. Atmos. Environ. 32: 1911-1919.
18 Yamasaki, H.; Kuwata, K.; Miyamoto, H. (1982) Effects of ambient temperature on aspects of airborne polycyclic
19 aromatic hydrocarbons. Environ. Sci. Technol. 16: 189-194.
20 Yatin, M.; Tuncel, S. G.; Tuncel, G.; Aras, N. K. (1994) Trace element composition of atmospheric aerosols in
21 Ankara, Turkey, determined by instrumental neutron activation analysis. J. Radioanal. Nucl. Chem.
22 181:401-411.
23 Zeng, X.; Wu, X.; Yao, H.; Yang, F.; Cahill, T. A. (1993) PIXE-induced XRF with transmission geometry.
24 Nucl. Instr. Meth. Phys. Res. B 75: 99-104.
25 Zhang, X. Q.; McMurry, P. H. (1987) Theoretical analysis of evaporative losses from impactor and filter deposits.
26 Atmos. Environ. 21: 1779-1789.
27 Zhang, X.; McMurry, P. H. (1992) Evaporative losses of fine paniculate nitrates during sampling. Atmos. Environ.
28 Part A 26: 3305-3312.
29 Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing characteristics and water content of
30 submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
31 27:1593-1607.
32 Zoller, W. H.; Gordon, G. E. (1970) Instrumental neutron activation analysis of atmospheric pollutants utilizing
33 Ge(Li) "gamma"-ray detectors. Anal. Chem. 42: 257-265.
34
April 2002 2-137 DRAFT-DO NOT QUOTE OR CITE
-------
i 3. CONCENTRATIONS, SOURCES, AND EMISSIONS
2 OF ATMOSPHERIC PARTICULATE MATTER
3
4
5 3.1 INTRODUCTION
6 This chapter discusses topics covered in Chapter 5 (Sources and Emissions of Atmospheric
7 Particles) and Chapter 6 (Environmental Concentrations) of the previous document, Air Quality
8 Criteria for Particulate Matter or " 1996 PM AQCD" (U. S. Environmental Protection Agency,
9 1996) and presents updates to these materials where available.
10 Information about concentrations, the composition, and the spatial and temporal variability
11 of ambient particles across the United States is presented in Section 3.2. Ambient concentration
12 data obtained during the first two years of operation of the recently deployed nationwide network
13 of Federal Reference Method PM2 5 monitors in twenty-seven metropolitan statistical areas
14 (MSAs) are presented and analyzed in Appendix 3A. Initial data from the pilot method
15 evaluation study for the national speciation network are presented in Appendix 3B. Results of
16 field studies that have characterized the composition of organic compounds in ambient particles
17 are summarized in Appendix 3C to complement the data for the inorganic composition of
18 ambient particles presented in Appendix 6A of the 1996 PM AQCD and Appendix 3B of this
19 document. Data for characterizing the daily and seasonal variability of PM25 concentrations are
20 discussed in Section 3.2.1, the intraday variability of PM25 concentrations in Section 3.2.2, the
21 relations among different size fractions in Section 3.2.3, the interrelations and correlations
22 among PM components in Section 3.2.4, and the spatial variability of various PM components in
23 Section 3.2.5.
24 Unlike gaseous criteria pollutants (SO2, NO2, CO, O3), which are well-defined chemical
25 entities, atmospheric parti culate matter (PM) is composed of a variety of particles differing in
26 size and chemical composition. Therefore, sources of each component of the atmospheric
27 aerosol must be considered in turn. Differences in the composition of particles emitted by
28 different sources also will lead to spatial and temporal heterogeneity in the composition of the
29 atmospheric aerosol. The nature of the sources and the composition of the emissions from these
30 sources are discussed in Section 3.3. The chemistry of formation of secondary PM from gaseous
April 2002 3-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 precursors is discussed in Section 3.3.1. The long-range transport of PM from sources outside
2 the United States is discussed in Section 3.3.2. Reviews of transport of PM and its precursors
3 within the United States can be found in the NARSTO Fine Particle Assessment (NARSTO,
4 2002). More detailed information regarding sulfur and nitrogen species can be found in Hidy
5 (1994). Estimates of contributions of various sources to ambient PM levels given by source
6 apportionment studies also are presented in Section 3.3.3. More detailed information about the
7 composition of emissions from various sources is given in Appendix 3D. Because PM is
8 composed of both primary and secondary constituents, emissions of both the primary
9 components and the gaseous precursors of secondary PM must be considered. Nationwide
10 emissions estimates of primary PM and precursors to secondary PM are discussed in
11 Section 3.3.4 and uncertainties in emissions estimates in Section 3.3.5.
12 The organization of topics in this chapter (ambient measurements, source characterization
13 and apportionment, and emissions inventories) reflects, in a broad sense, the order in which these
14 topics are addressed in scientific studies and, arguably, the increasing levels of uncertainty that
15 are associated with these topics.
16
17
18 3.2 PATTERNS AND TRENDS IN AMBIENT PM CONCENTRATIONS
19 A significant amount of data for characterizing PM10 mass concentrations and trends exists,
20 and that available up to about 1994 was presented in the 1996 PM AQCD. However, data sets
21 for characterizing PM2 5 and PM10_2 5 mass or trends were not as extensive. Sources of data for
22 PM25 (fine) and PM10_2 5 (coarse), which were discussed in the 1996 PM AQCD, include EPA's
23 Aerometric Information Retrieval System (AIRS) (U.S. Environmental Protection Agency,
24 2000a), IMPROVE (Eldred and Cahill, 1994; Cahill, 1996), the California Air Resources Board
25 (CARS) Data Base (California Air Resources Board, 1995), the Harvard Six-Cities Data Base
26 (Spengler et al., 1986; Neas, 1996), and the Harvard Philadelphia Data Base (Koutrakis, 1995).
27 The Inhalable Particulate Network (IPN) (Inhalable Paniculate Network, 1985; Rodes and Evans,
28 1985) provided TSP, PM15, and PM2 5 data but only a small amount of PM10 data.
29 New sources of PM data include the recently deployed nationwide PM25 compliance
30 monitoring network, which provides mass measurements using a Federal Reference Method
31 (FRM). This section summarizes data obtained during 1999 and 2000 by this network and
April 2002 3-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 provides an approximate characterization of nationwide PM10_25 concentrations by comparing
2 PM10 to PM2 5 measurements at sites where both types of compliance monitors are located.
3 Various aspects of these data are presented in greater detail in Appendix 3 A. In addition, a small
4 number of recent studies in which daily mass and composition measurements are available for
5 extended periods are discussed in this section. The results of quality assured aerosol composition
6 data obtained by X-ray fluorescence (XRF) and by analyses of organic carbon (OC) and
7 elemental carbon (EC) for thirteen urban areas from the methods evaluation study for the national
8 PM25 speciation network are presented in Appendix 3B.
9 Organic compounds contribute from 10 to 70% of the dry fine particle mass in the
10 atmosphere (see Appendix 3C). However, organic PM concentrations, composition, and
11 formation mechanisms are poorly understood. Particulate organic matter is an aggregate of
12 hundreds of individual compounds spanning a wide range of chemical and thermodynamic
13 properties (Saxena and Hildemann, 1996). Some of the organic compounds are "semivolatile"
14 (i.e., they have atmospheric concentrations and saturation vapor pressures such that both gaseous
15 and condensed phases exist in equilibrium in the atmosphere). The presence of semivolatile or
16 multiphase organic compounds complicates the sampling process. Organic compounds
17 originally in the gas phase may be absorbed on glass or quartz filter fibers and create a positive
18 artifact. Conversely, semivolatile compounds originally present in the condensed phase may
19 evaporate from particles collected on glass, quartz, or Teflon filters creating a negative artifact.
20 In addition, no single analytical technique is currently capable of analyzing the entire range of
21 organic compounds present in atmospheric PM. Rigorous analytical methods are able to identify
22 only 10 to 20% of the organic PM mass on the molecular level (Rogge et al., 1993), and only
23 about 50% of the condensed phase compounds could be identified in smog chamber studies of
24 specific compounds (Forstner et al., 1997a,b). Measurement techniques are discussed in
25 Section 2.2.3.2. Information on the identification and concentration of the many different
26 organic compounds identified in atmospheric samples obtained during the 1990s is given in
27 Appendix 3C.
28 Summary tables giving the results of 66 field studies that obtained data for the composition
29 of particles in the PM2 5, PM10_25, or PM10 size ranges were presented in Appendix 6 A of the 1996
30 PM AQCD. The summary tables include data for mass, organic carbon, elemental carbon,
31 nitrate, sulfate, and trace elements. Data from the studies were presented for the eastern, western,
April 2002 3-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 and central United States. It should be noted that these studies took place at various times and
2 lasted for various durations over a 20-year period, and there may have been significant changes in
3 the concentrations of many species between the times when these studies were conducted and
4 now. These changes resulted from a number of factors (e.g., pollution controls, technological
5 advances, land use changes, etc).
6 There were a number of discernible differences in the composition of particles across the
7 United States evident in the data sets listed in Appendix 6A in the 1996 PM AQCD (cf
8 Figures 6-85a to 6-85c). However, these differences can only be discussed in the context of the
9 uncertainties in the measurements of the main components (sulfate, organic carbon, elemental
10 carbon, crustal material, ammonium, and nitrate). Sulfate, followed by crustal materials have the
11 smallest uncertainties associated with their measurement among all the components listed.
12 Sulfate constituted about 38% of PM2 5 in the aerosol composition studies in the eastern United
13 States and was the major identifiable component of PM2 5, but it constituted only about 11% of
14 PM2 5 in the studies listed for the western United States. The contribution of crustal materials to
15 PM2 5 ranged from about 4% in the East to about 15% in the West. The contribution of
16 unidentified material (possibly consisting mostly of water of hydration) ranged from 23% in the
17 East to 0% in the West. The contribution of elemental carbon to PM2 5 ranged from about 4% in
18 the East to about 15% in the West. Organic compounds constituted about 21% of PM2 5 in the
19 eastern United States, ranging to about 39% for the studies listed in the western United States.
20 However, uncertainties for organic carbon, elemental carbon, ammonium, and nitrate are larger
21 than for sulfate and crustal material. A factor of 1.4 was used to account for the presence of
22 oxygen and nitrogen in the organic compounds. This factor may vary among different areas and
23 may represent the lowest reasonable estimate for an urban aerosol (Turpin and Lim, 2001). In
24 addition, the samples collected in the studies were subject to a number of sampling artifacts
25 involving the adsorption of gases and the evaporation of volatile components that either formed
26 on the filters or were present in the ambient particles. The values reported for organic carbon and
27 elemental carbon in filter samples depend strongly on the specific analysis method used (Chow
28 etal., 2001).
29 Crustal materials constitute from 52% of PM10_2 5 in the eastern United States to 70% of
30 PM10_2 5 in the studies in the western United States given in Appendix 6A of CD96. The fraction
31 of unidentified material in PM10_2 5 varied from 41% in the eastern United States to 27% in the
April 2002 3-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 western United States. However, in the vast majority of these studies no attempt was made to
2 characterize organic components or nitrate in the PM10_25 size fraction. Even if analyses of total
3 OC were available, they would not be able to distinguish between bioaerosols and simple organic
4 compounds. Indeed, in many photomicrographs of PM10_25 samples obtained by scanning
5 electron microscopy, the fields of view were dominated by large numbers of pollens, plant and
6 insect fragments, and microorganisms. Bioaerosols such as pollens, fungal spores, and most
7 bacteria are expected to be found mainly in the coarse size fraction. However, allergens from
8 pollens can also be found in fine particles (Monn, 2001). It should also be remembered that a
9 small fraction (typically about 10%) of PM2 5 is entrained into the flow of the channel of the
10 dichotomous sampler that collects the PM10_25 sample and that there may be errors invoked
11 during the procedure used to account for this entrainment.
12 Data for the chemical composition of particles in a number of national parks and remote
13 areas have been collected for a number of years by the IMPROVE network. Concentrations are
14 reported for sulfate, nitrate, light absorbing carbon, organic carbon, and soil components. With
15 the collection of compositional data by the speciation network, more synoptic (i.e., concurrent)
16 coverage will be obtained for these constituents in continental background to urban environments
17 across the United States.
18
19 PM10 Concentrations and Trends
20 Nationwide PM10 annual mean concentrations on a county-wide basis from the AIRS
21 database for calendar years 1999 and 2000 are shown in Figure 3-1 a. Concentrations in most
22 areas of the country were below the level of the PM10 annual standard (50 //g/m3) in 1999 and
23 2000. The median annual PM10 concentration was about 23 //g/m3; and five percent of the
24 countywide concentrations shown in Figure 3-la were greater than 35 //g/m3. The 98th percentile
25 PM10 concentrations are shown in Figure 3-lb. Data from all monitors for the most recently
26 available four consecutive quarters in 1999 and 2000 with at least eleven valid observations per
27 quarter in a given county were averaged to produce Figure 3-la; and data from the highest
28 monitor in that county were used to produce Figure 3-lb. In these, and similar maps for PM2 5
29 and PM10_2 5, cut points were chosen at the median and 95th percentile concentrations. As shown
30 by the blank areas on the maps, the picture is not complete because some monitoring locations
31 did not record valid data for all four quarters or recorded fewer than 11 samples in one or more
April 2002 3-5 DRAFT-DO NOT QUOTE OR CITE
-------
Virgin Islands
Alaska
Concentration ((jg/m3)
Hawaii
"' 0 < x <= 23
Puerto Rico
23 < x <= 35
x>35
Figure 3-la. 1999-2000 county-wide average annual mean PM10
concentrations
Source: U.S. EPA Aerometric Information Retrieval System (12/19/2001).
.,•**
Virgin Islands
Alaska
Concentration (Mg/m3)
Hawaii
Puerto Rico
0115
Figure 3-lb. 1999-2000 highest county-wide 98th percentile 24-h average
PM10 concentrations
Source: U.S. EPA Aerometric Information Retrieval System (12/19/2001).
April 2002 3-6 DRAFT-DO NOT QUOTE OR CITE
-------
1 quarters or counties simply did not have monitors. Similar considerations apply to the maps to
2 be shown later for PM2 5 and PM10_2 5. It should also be noted that the area of counties can be
3 much greater in the West than in the East. As a result, the density of monitors may appear to be
4 greater in the West and air quality may appear to be worse over much larger areas in the West
5 than in the East.
6 Nationwide trends in annual mean PM10 concentrations from 1990 through 1999 (based on
7 data obtained at 153 rural sites, 375 suburban sites, and 408 urban sites reporting to AIRS) are
8 shown in Figure 3-2 (U.S. Environmental Protection Agency, 2001). Though average
9 concentration levels differ among sites, with higher levels at urban and suburban sites, the
10 nationwide data set shows a decrease of 18% that occurred mainly during the first half of the
11 record. PM10 concentrations then leveled off during the last few years of the record. Figure 3-3
12 shows the annual mean PM10 trend summarized by EPA region. Decreases in annual average
13 PM10 concentrations from 1990 to 1999 were largest in the Northwest (10.3 //g/m3) and smallest
14 in the south central United States (3.2 //g/m3). Analyses of available TSP measurements
15 obtained since 1950 indicate that mean TSP concentrations could have declined two- to three-
16 fold in urban areas between 1950 and 1980 (Lipfert, 1998).
17
18 PM25 Concentrations and Trends
19 Nationwide annual mean PM2 5 concentrations for 1999 and 2000 are shown in Figure 3-4a
20 and 98th percentile concentrations are shown in Figure 3-4b. Quantities shown in Figure 3-4a and
21 3-4b were calculated for individual counties. Data from all monitors in a given county meeting
22 the same minimum data completeness criteria for PM10 (given earlier) were averaged to produce
23 Figure 3-4a, and results from the highest monitor were used to produce Figure 3-4b. The median
24 PM25 concentration nationwide was about 13 //g/m3. Annual mean PM25 concentrations were
25 above 18 //g/m3 at 5% of the sites, mainly in California and in the southeastern United States.
26 The 98th percentile 24-h average concentrations were below 50 //g/m3 at 95% of the sites
27 sampled. Most of the sites with levels above 50 //g/m3 are located in California.
28 Annual average PM25 concentrations obtained as part of health studies conducted in
29 various locations in the United States and Canada from the late 1980s to the early 1990s are
30 shown in Figure 3-5 (Bahadori et al., 2000a). These studies include the Harvard six-cities study
(Steubenville, OH; Watertown, MA; Portage, WI; Topeka, KS; St. Louis, MO; and Kingston-
April 2002 3-7 DRAFT-DO NOT QUOTE OR CITE
-------
35
30
25
20
15
10
Rural (153 sites) Suburban (375 sites) Urban (408 sites)
90 91 92 93 94 95 96 97 98 99
Figure 3-2. Nationwide trend in ambient PM10 concentration from 1990 through 1999.
Source: U.S. Environmental Protection Agency (2001).
Trend in PM10 annual mean concentration by EPA Region, 1989-1998.
Alaska is in EPA Region 10; Hawaii, EPA Region 9; and Puerto Rico, EPA Region 2.
Concentrations are ug/m:i.
Note: These trends are
influenced by the
distribution of monitoring
locations in a given region
and, therefore, can be
driven largely by urban
concentrations. For this
reason, they are not
indicative of background
regional concentrations.
Figure 3-3. Trend in PM10 annual mean concentrations by EPA region, 1990 through
1999
Source: U. S. Environmental Protection Agency (2001).
April 2002
3-8
DRAFT-DO NOT QUOTE OR CITE
-------
>"
* **"* **
>
Virgin Islands
Alaska
Concentration (Mg/m3)
Hawaii
^~ 0
** ^ 5* '
Virgin Islands
Alaska
Concentration (Mg/m3)
Hawaii
' 0 < x <= 33
Puerto Rico
Figure 3-4b. 1999-2000 highest county-wide 98th percentile 24-h
average PM2 5 concentrations
Source: U.S. EPA Aerometric Information Retrieval System (12/19/2001).
April 2002 3-9 DRAFT-DO NOT QUOTE OR CITE
-------
>
to
o
o
to
H
6
o
2
o
H
O
o
?d
o
c^ 140-
E
§ 120-
s — '
c
.2 100-
CO
^
§ 8°-
o
c
0 60-
O
CD
TO 40-
0
'•E 20-
03
Q_
0-
• 95%
PM2 5 Annual otandard T 90%
o, JL 750/.
1 1 ^ i in/m i
l I \J Uy/ III y
*^ ~ Mean
- 50%
LJ 25%
1 10%
•
•
• 5%
•
I ' ' • T • ' ' « ' "
' JL T T T ' . T* • * T " T T ' f '
-Pir^iJL^r1!!^^ Illlll T T f T I T ••••'••
] - a a HHHggQgi^iiiMi^
• * • ^T . TT T
City
Figure 3-5. Collection of annual distribution of 24-h average PM2 5 concentrations observed in U.S.
and Canadian health studies conducted during the 1980's and early 1990's.
Source: Bahadori et al. (2000a).
-------
1 Harriman, TN); PTEAMS (Riverside, CA); MAACS (Philadelphia, PA; Washington, DC; and
2 Nashville, TN); South Boston Air Quality and Source Apportionment Study (Boston, MA); and
3 NPMRMN (Phoenix, AZ). The remaining sites were part of the 24-cities study (Spengler et al.,
4 1996).
5 Sufficient data are not yet available to permit the calculation of nationwide trends of PM2 5
6 and PM10_25; however, some general conclusions can be reached. Darlington et al. (1997)
7 proposed that the consistent reductions in PM10 concentrations found in a wide variety of
8 environments ranging from urban to rural may have resulted from common factors or controls
9 that affected fine particles more strongly than coarse particles. This is because fine particles have
10 longer atmospheric lifetimes than coarse particles and can be transported over longer distances
11 and, hence, can affect larger areas. Apart from the IMPROVE network of monitoring sites
12 located mainly in national parks, the longest time series of PM25 concentration and composition
13 data have been obtained by the California Air Resources Board (CARB). Their data show that
14 annual average PM25 concentrations decreased by about 50% in the South Coast Air Basin, 35%
15 in the San Joaquin Valley, 30% in the San Francisco Bay Area, and 35% in the Sacramento
16 Valley from 1990 to 1995 (Dolislager and Motallebi, 1999). PM2 5 data were collected
17 continuously from 1994 to 1998 as part of the children's health study in 12 communities in
18 southern California (Taylor et al., 1998). Data obtained at all sites show decreases in PM25
19 ranging from 2% at Santa Maria to 37% at San Dimas/Glendora from 1994 through 1998. These
20 decreases were accompanied by decreases in major components such as nitrate, sulfate,
21 ammonium, and acids. Based on the analysis of PM2 5 data sets collected prior to 1990, Lipfert
22 (1998) found that PM25 concentrations could have decreased by about 5% per year from 1970 to
23 1990 in a number of urban areas. These declines were also found to be consistent with decreases
24 in emissions from combustion sources over that time period.
25
26 Background PM2 5 Concentrations
27 In common usage, the term "background concentrations" refers to concentrations observed
28 in remote areas relatively unaffected by local pollution sources. However, as noted in Chapter 6
29 of the 1996 PM AQCD, several definitions of background concentrations are possible. In that
30 document, the two definitions chosen as being most relevant for regulatory purposes are based on
31 estimates of contributions from uncontrollable sources that can affect concentrations in the
April 2002 3-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 United States. The first definition refers to the concentration resulting from anthropogenic and
2 natural emissions outside North America and natural sources within North America. The second
3 definition refers to the concentration resulting from natural sources only within and outside of
4 North America. Because of long-range transport from anthropogenic source regions in North
5 America, it is impossible to obtain background concentrations as defined above solely on the
6 basis of direct measurement in remote areas in North America. However, these data can be used
7 to place reasonable upper limits on what these concentrations could be. The range of values in
8 the lowest 5th percentile annual mean PM2 5 concentrations in the AIRS data base is from
9 2.8 //g/m3 to 6.9 //g/m3. This range of concentrations is consistent with the range of annual mean
10 PM25 concentrations at remote sites in the western United States obtained from 1996 through
11 1999 in the IMPROVE network. At most IMPROVE sites in the western United States, the
12 mean concentration of PM10_25 is higher than that of PM25, and PM2 5 concentrations are
13 moderately correlated (r = 0.72) with PM10_2 5 concentrations. In contrast, PM2 5 concentrations
14 are higher than those of PM10_2 5 at IMPROVE sites in the eastern United States, and PM2 5
15 concentrations are only weakly correlated (r = 0.26) with those of PM10_25.
16 Annual average natural background concentrations of PM10 (according to definition 1) have
17 been estimated to range from 4 to 8 //g/m3 in the western United States and 5 to 11 //g/m3 in the
18 eastern United States. Corresponding PM25 levels have been estimated to range from 1 to
19 4 //g/m3 in the western United States and from 2 to 5 //g/m3 in the eastern United States (U.S.
20 Environmental Protection Agency, 1996). Although these values are broadly consistent with the
21 data given above, the data discussed in the previous paragraph represent only upper limits to
22 background concentrations because of possible contributions from long-range transport from
23 anthropogenic sources within North America. Peak 24-h average natural background
24 concentrations may be substantially higher than the annual or seasonal average natural
25 background concentrations, especially within areas affected by wildfires and dust storms and
26 long range transport from outside North America. Estimates of background concentrations
27 according to definition 2 are not yet available. However, recent information about contributions
28 to background concentrations that fall under definitions 1 and 2 because of long-range transport
29 from sources outside the United States is given in Section 3.3.2.
30
31
April 2002 3-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 PM10_2,5 Concentrations
2 By using AIRS data for 1999 and 2000 obtained by the PM10 and PM25 compliance
3 networks, it is possible to construct a picture of the distribution of coarse PM across the country.
4 This is accomplished by pairing data from 228 compliance monitoring sites where PM10 and
5 PM25 monitors are collocated and subtracting the mass concentrations of PM2 5 from PM10.
6 Nationwide annual mean PM10_2 5 concentrations calculated by this difference method are shown
7 in Figure 3-6a. Annual mean PM10_25 concentrations range from 1 to 48 //g/m3, with a
8 nationwide median concentration of about 10 //g/m3; and 5% of the sites had mean
9 concentrations greater than 20 //g/m3. The higher values occur mainly in the western United
10 States, particularly in California. The highest county-wide 98th percentile PM10_2 5 concentrations
11 based on this same data set are shown in Figure 3-6b. Highest values in the western United
12 States are caused by dust raised locally either by natural means or by anthropogenic activity.
13 Elevated dust levels are also found in southern Florida as the result of dust storms in North
14 Africa (cf Section 3.3.2). In many areas, combined errors in the PM25 and PM10 measurements
15 may be similar to or even greater than PM10_25 concentrations. Because of this and other
16 potential problems with this approach (cf. Section 3.2.1), these results should be viewed with
17 caution.
18
19 3.2.1 Seasonal Variability in PM Concentrations
20 PM25
21 Aspects of the spatial and temporal variability of PM25 concentrations for 1999 and 2000 in
22 a number of metropolitan areas across the United States are presented in this and following
23 subsections. Data for multiple sites in 27 urban areas across the United States have been
24 obtained from the AIRS data base and analyzed for their seasonal variations and for their spatial
25 correlations and spatial uniformity in concentrations (Pinto, et al., 2002). Only 27 MS As were
26 included in the analyses based on the criteria that data be obtained on at least 15 days in each
27 calendar quarter of either 1999 and 2000 or 2000 alone at four sites within that MSA. A number
28 of aspects of the spatial and temporal variability of the 1999 PM25 data set were presented in
29 Rizzo and Pinto (2001), based in part on analyses given in Fitz-Simons et al. (2000).
30
April 2002 3-13 DRAFT-DO NOT QUOTE OR CITE
-------
LJ
Virgin Islands
Alaska
Concentration
Hawaii
" i 020
Figure 3-6a. 1999-2000 estimated county-wide average annual mean PM10_2 5
concentrations
Source: U.S. EPA Aerometric Information Retrieval System (12/19/2001).
Virgin Islands
Alaska
Concentration (|jg/m3)
Hawaii
-•" I 070
Figure 3-6b. 1999-2000 estimated county-wide highest 98th percentile 24-h average PM10_2 5
concentrations
DRAFT-DO NOT QUOTE OR CITE
Source: U.S. EPA Aerometric Information Retrieval System.
April 2002 3-14
-------
1 Information regarding the seasonal variability in PM2 5 concentrations in four MSAs
2 (Columbia, SC; Detroit, MI; Chicago, IL; Los Angeles-Long Beach, CA) in the United States is
3 summarized in Figures 3-7a through 3-7d. These four urban areas were chosen to illustrate some
4 general features of the spatial and temporal variability found in the United States. The figures
5 show lowest, lower quartile, median, upper quartile, and highest concentrations for each calendar
6 quarter of 1999 and 2000 for the Columbia, SC, and Los Angeles, CA MSAs and for 2000 for
7 the Detroit, MI and Chicago, IL MSAs. For each monitoring site, the AIRS ID numbers, annual
8 mean concentrations, the number of observations, and standard deviations are also shown. Data
9 for multiple sites within these MSAs are shown to provide an indication of the degree of inter-
10 site variability. Data for these MSAs and an additional twenty-three MSAs, criteria used for site
11 selection, and additional descriptions of the data are given in Appendix 3 A.
12 Annual mean PM25 concentrations (based on two years data) at individual monitoring sites
13 in the MSAs examined range from about 6 //g/m3 to about 30 //g/m3. The lowest values are
14 found in rural portions of the MSAs examined, typically near the perimeter of the MSA. Annual
15 mean concentrations tend to be higher in the Southeast than in the Northeast and higher in
16 southern California compared to the Pacific Northwest (cf Appendix 3 A). However, average
17 PM25 concentrations tend to be lower in 1999 and 2000 in urban areas given in Appendix 3 A
18 compared to the concentrations observed during pollution-health outcome studies conducted in
19 those five urban areas where these overlap (cf. Figure 3-5). It should be noted that there are no
20 data demonstrating the comparability of the monitors used in the studies shown in Figure 3-5 and
21 theFRM.
22 In four of the seven MSAs examined in the eastern United States (as in the Columbia, SC
23 MSA, cf. Figure 3-7a), highest median concentrations occur at most sites during the third
24 calendar quarter (i.e., summer months). There are exceptions to this pattern as shown for the
25 Philadelphia, PA-NJ MSA (cf. Figure 3A-1). Highest median concentrations in the north-central
26 United States tend to occur in the first or fourth quarters (i.e., winter months) as in the Detroit,
27 MI and Chicago, IL MSA (cf. Figures 3-7b and 3-7c). Highest median concentrations occur
28 during the fourth calendar quarter in MSAs in the western United States as in the Los Angeles,
29 CA PMSA (cf. Figure 3-7d), although there are exceptions at individual sites in the Riverside,
30 CA PMSA.
31
April 2002 3-15 DRAFT-DO NOT QUOTE OR CITE
-------
AIRS \D#
Mean
Obs
SD
60
50
40
30
20
10
A. Columbia, SC MSA
4506300051 4506300081 4507900071 4507900191
14.680
231
6.760
16.462
228
7.121
15.461
216
6.900
16.098
229
7.148
1234 1234 1
B. Detroit, Ml MSA
AIRSIDtf
Mean
Obs
SD
50 •
40 •
30 -
20 •
10 •
0 •
2609900091 2612500011 2614700051 2616300331 2616300361
13.450
113
7.922
I
I
I
I
i
1 5.552
90
9.223
I
,
I
N
I I
14.172
102
8.771
||
'!
I
20.173 17.446
1 08 1 03
10.475 9.626
I;
I
I
I
I!
I
I
Figure 3-7a,b. Quarterly distribution of 24-h average PM2 5 concentrations for selected
monitors in the (a) Columbia, SC; (b) Detroit, MI; (c) Chicago, IL; and
(d) Los Angeles, CA MSAs. Values for the lowest, lower quartile, median,
upper quartile and highest concentrations are shown in the figures. The
AIRS site ID number, annual concentration, number and standard
deviation are shown above the figures for each site.
April 2002
3-16
DRAFT-DO NOT QUOTE OR CITE
-------
C. Chicago, IL MSA
AIRSID* 1703100141 1703100221 1703100501 1703100521 1703110161 1703120011 1703133011 1703140061 1703142011 1704340021 1719710021
Mean
Obs
SD
15.823
104
7.935
17.933
113
8.175
16.996
274
8.468
18.295
346
9.289
20.277
108
9.331
16.790
113
7.694
16.889
115
7.689
15.268
101
8.423
14.283
327
7.905
15.215
116
7.568
15.994
112
7.405
I,
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
D. Los Angeles
AIRS IDS 0603700021 0603711031 0603716011 0603740021 0603790021
Mean
Obs
SD
100 •
75 •
50 •
25 •
0 •
21.682
469
13.923
,
„
II
i
1
i
I
22.207 24.764 20.225 10.917
428 218 417 204
13.840 14.056 12.994 5.043
|,
I I
'!
Hi
,
Hi
,
r
W
1
|
I
Mill
1234 12
1234 1234
Figure 3-7c,d. Quarterly distribution of 24-h average PM2 5 concentrations for selected
monitors in the (a) Columbia, SC; (b) Detroit, MI; (c) Chicago, IL; and
(d) Los Angeles, CA MSAs. Values for the lowest, lower quartile, median,
upper quartile and highest concentrations are shown in the figures. The
AIRS site ID number, annual concentration, number and standard
deviation are shown above the figures for each site.
Source: Pinto et al. (2002).
April 2002
3-17
DRAFT-DO NOT QUOTE OR CITE
-------
1 Lowest median concentrations occur mainly during the first or fourth quarters at most sites
2 in the eastern United States, with some occurring during the second quarter (cf. Appendix 3 A).
3 In moving westward, the seasonal pattern is not as distinct, with lowest median concentrations
4 occurring in any quarter, but usually in the second or third quarter as in the Chicago, IL and
5 Detroit, MI MS As (cf. Figure 3-7b and 3-7 c). In many of the MS As examined, seasonal
6 variations follow a similar pattern at all of the sites within the MSA, but in other MS As there are
7 noticeable differences in the seasonal pattern between sites. The large-scale differences in
8 seasonal variability between MS As tend to follow differences in the major categories of PM
9 sources affecting the monitoring sites. Local heating by wood burning during the colder months
10 is practiced more widely in the western United States than in the eastern United States. Hence,
11 winter maxima and greater variability in PM2 5 concentrations across sites are expected in the
12 West due to the influence of the local sources. On the other hand, photochemical production of
13 secondary PM, especially sulfate, occurs over wide areas in relatively homogeneous air masses
14 during the summer months in the eastern United States. Because sulfates (along with associated
15 cations and water) constitute the major fraction of summertime PM2 5 in the East, there is greater
16 uniformity in 3rd quarter PM concentrations within eastern MSAs (cf. Appendix 3 A).
17 The highest values shown in the box plots in Figures 3-7a to 3-7d and in Figures 3A-1 to
18 3A-27 do not always follow the same seasonal pattern as do the median concentrations. These
19 values likely reflect the existence of transient events such as forest fires (mainly in the West) or
20 episodes of secondary PM production (mainly in the East). However, chemical analyses of filter
21 samples or other evidence should be used to determine specific causes in particular locations.
22 There have been a few studies that have characterized PM2 5 and PM10 concentrations in
23 major urban areas. The Metropolitan Acid Aerosol Characterization Study (MAACS) (Bahadori
24 et al., 2000b) characterized the levels and the spatial and temporal variability of PM2 5, PM10, and
25 acidic sulfate concentrations in four cities in the eastern United States (Philadelphia, PA;
26 Washington, D.C.; Nashville, TN; and Boston, MA). Seasonal variations in PM2 5 and PM10
27 concentrations obtained during the course of this study are shown in Figure 3-8. The data for the
28 four cities included in MAACS are presented as box plots showing the lowest, lowest tenth
29 percentile, lowest quartile, median, highest quartile, highest tenth percentile, and highest PM2 5
30 and PM10 values. Mean and highest PM2 5 and PM10 concentrations are found during the summer
31 in all four cities, although the seasonal pattern in Boston appears to be more nearly bimodal with
April 2002 3-18 DRAFT-DO NOT QUOTE OR CITE
-------
g
"ro
cu
o
c
o
O
g
'•c
03
D_
70
60-
50-
40
30-
20-
10-
0
Philadelphia | Washington
Boston
SPSU F W SPSU F W SPSU F W SPSU F W
Season
Figure 3-8. Concentrations of PM2 5 and PM10 measured in the four MAACS cities. The
data show the lowest, lowest tenth percentile, lowest quartile, median highest
quartile, highest tenth percentile, and highest PM2 5 24-h average values. The
dashed line shows the level of the annual PM2 5 standard.
Source: Bahadori et al. (2000b).
1
2
3
4
5
6
7
8
9
10
11
12
13
an additional winter peak. This seasonal pattern, based on 2- to 3-year sampling periods for each
city during 1992 through 1996, is in accord with that obtained from the FRM monitors in the
NAMS and SLAMS network (cf Appendix 3A).
PM
10-2.5
Data from the FRM PM2 5 and PM10 compliance networks that could be used to characterize
seasonal variations in PM10_2 5 based on EPA minimum data completeness criteria (11 samples
per calendar quarter) are available for 228 sites nationwide. Data for the seasonal variations in
PM10_2 5 concentrations for Columbia, SC and Detroit, MI are shown for 2000 and data for Los
Angeles-Long Beach are shown for 1999 in Figures 3-9a,b,c. As can be seen by comparing the
number of observations used in the calculation of values shown in Figures 3-7a,b,c,d and Figure
3-9a,b,c the number of days that could be used for calculating PM10_25 concentrations is much
less than that measured for PM2 5. At least for the sites shown for Columbia, SC; and Detroit,
April 2002
3-19
DRAFT-DO NOT QUOTE OR CITE
-------
Columbia, SC MSA
(a)
AIRSID*
Mean
Obs
SD
40 •
30 •
20 •
10 •
0 •
450790019
9.643
53
6.172
„
II
i
1
(b)
Mean
Obs
SD
Detroit, Ml MSA
AIRSID* 261630001 261630015 261630025
80 1
60
40 •
20 •
-20 •
11.517
56
10.262
19.416
58
15.611
7.328
55
7.638
Figure 3-9a,b. Quarterly distribution of 24-h average PM10_2 5 concentrations for selected
sites in the (a) Columbia, SC; (b) Detroit, MI; and (c) Los Angeles, CA
MSAs. Values for the lowest, lower quartile, median, upper quartile and
highest concentrations are shown in the figures. The AIRS site ID number,
annual concentration, number and standard deviation are shown above the
figures for each site.
April 2002
3-20
DRAFT-DO NOT QUOTE OR CITE
-------
(c)
Los Angeles, CA MSA
AIRSIDtf
Mean
Obs
SD
0603700021 0603711031 0603716011 0603740021 0603790021
100 -
75 -
50 -
25 -
0 H
21.682 22.207 24.764 20.225 10.917
469 428 218 417 204
13.923 13.840 14.056 12.994 5.043
I
I
II
,
,
„
,!
I
li in
i
,
!
!M ltt|
1234
1234
1234
1234
Figure 3-9c. Quarterly distribution of 24-h average PM10_2 5 concentrations for selected
sites in the (a) Columbia, SC; (b) Detroit, MI; and (c) Los Angeles,
CA MSAs. Values for the lowest, lower quartile, median, upper quartile
and highest concentrations are shown in the figures. The AIRS site ID
number, annual concentration, number and standard deviation are shown
above the figures for each site.
Source: Pinto et al. (2002).
1 MI the seasonal median maxima in concentrations occur during the second calendar quarter.
2 It can readily be seen that a number of PM10_25 concentrations are negative. (The negative
3 estimates have been included in the calculation of mean concentrations.) There are a number of
4 reasons for the negative concentration estimates, many of which arise because the ratios of PM25
5 to PM10 are based on two independent measurements. Measurement imprecision plays a role
6 when the ratios are large and concentrations are small. Differences in the behavior of
7 semivolatile components in the two samplers could occur. The results may be due to errors in
April 2002
3-21
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
9
10
11
12
13
14
15
16
sampler placement, field, laboratory, or data processing procedures. Therefore, caution should be
exercised when attempting to interpret results for PM10_2 5 based on the current network
collocated PM2 5 and PM10 monitors.
Frequency Distributions for PM25 Data
Frequency distributions for PM2 5 concentrations obtained in Philadelphia from 1992
through 1995 are shown in Figure 3-10 (data obtained by Bahadori et al., 2000b). Concentrations
predicted from the log-normal distribution, using geometric mean values and standard deviations
derived from the data, are also shown. In Philadelphia, the highest PM2 5 values were observed
when winds were from the southwest during sunny but hazy high pressure conditions.
In contrast, the lowest values were found after significant rainstorms during all seasons of the
year. Mean ± SD day-to-day concentration differences in the data set are 6.8 ± 6.5 //g/m3 for
PM25 and 8.6 ± 7.5 //g/m3 for PM10. Maximum day-to-day concentration differences are
54.7 Mg/m3 for PM2 5 and 50.4 //g/m3 for PM10.
350
PM25
geometric mean = 15.2 |jg/m3
a =1.69
Concentration (|jg/m )
Figure 3-10. Frequency distribution of 24-h average PM2 5 concentrations measured at the
Presbyterian home (PBY) monitoring site in southwestern Philadelphia from
1992 to 1995. Log-normal distribution fit to the data shown as solid line.
April 2002
3-22
DRAFT-DO NOT QUOTE OR CITE
-------
1 Different patterns are observed in data collected elsewhere in the United States. PM2 5
2 concentrations obtained in Phoenix, AZ from 1995 through 1997 (Zweidinger et al., 1998) are
3 summarized in Figure 3-11, and frequency distributions of PM2 5 concentrations obtained in
4 Phoenix are shown in Figure 3-12. Mean ±SD day-to-day concentration differences in this data
5 set are 2.9 ±3.0 //g/m3 with a maximum day-to-day concentration difference of 23 //g/m3. PM25
6 and PM10_2 5 data were obtained with dichotomous samplers at a number of sites in California on
7 a sampling schedule of every 6 days from 1989 through 1998. Histograms showing the
8 frequency distribution of the entire set of PM25 and PM10_25 concentrations obtained by the
9 CARB network of dichotomous samplers from 1989 to 1998 are shown in Figures 3-13 and 3-14.
10 Also shown are log-normal distributions generated using geometric means and standard
11 deviations derived from the data as input. Although the data for both size fractions appear to be
12 reasonably well simulated by the function, data obtained at individual locations may not be. Data
13 showing the seasonal variability of PM25 obtained at Riverside-Rubidoux are summarized in box
14 plot form in Figure 3-15. The frequency distribution of PM25 concentrations obtained at
15 Riverside-Rubidoux from 1989 to 1994 is shown in Figure 3-16. It can be seen that the data are
16 not as well fit by a log-normal distribution as are the data shown in Figure 3-10, partially as the
17 result of a significant number of days when PM2 5 concentrations are greater than 100 //g/m3.
18 An examination of the data from the four MAACS cities, Phoenix, AZ, and Riverside, CA,
19 indicates that substantial differences exist in aerosol properties between the eastern cities
20 (MAACS) and the western cities (Phoenix, AZ; Riverside, CA). Fine-mode particles account for
21 most of the PM10 mass observed in the MAACS cities and appear to drive the daily and seasonal
22 variability in PM10 concentrations there. Coarse-mode particles represent a larger fraction of
23 PM10 mass in Phoenix and Riverside and drive the seasonal variability in PM10 seen there. The
24 average ratio of PM2 5 to PM10 concentrations is much larger in the MAACS cities of
25 Philadelphia (0.72); Washington, DC (0.74); and Nashville (0.63) than in either Phoenix (0.34)
26 or Riverside (0.49). Differences between median and maximum concentrations in any size
27 fraction are much larger at the Riverside site than at either the MAACS or Phoenix sites. Many
28 of these differences could reflect the more sporadic nature of dust suspension at Riverside.
29 In addition, the seasonal variability of PM25 concentrations observed in Phoenix, AZ, and
30 Riverside, CA, appears to be different from that observed in the MAACS cities. These
April 2002 3-23 DRAFT-DO NOT QUOTE OR CITE
-------
tu
£ 30-
n.
o 20-
"ro
c
0
o ..^
c 10-
o
O
n -
Phoenix, AZ T
I I
| |
V
PM25
(n = 876)
I |
0
¥
r
1
I |
•
•
Y V
Mar - May June-Aug Sep-Nov Dec-Feb
Figure 3-11. Concentrations of 24-h average PM25 measured at the EPA site in Phoenix,
AZ from 1995 to 1997. The data show the lowest, lowest tenth percentile,
lowest quartile, median (black circles), highest quartile, highest tenth
percentile, and highest PM2 5 values.
Source: Zweidingeretal. (1998).
200
o
o
PM2.5
geometric mean = 10.5 |jg/m3
= 1.70
10 15 20 25
Concentration (|jg/m3)
35 40
Figure 3-12. Frequency distribution of 24-h average PM2 5 concentrations measured
at the EPA site in Phoenix, AZ from 1995 to 1997.
Source: Zweidingeretal. (1998).
April 2002
3-24
DRAFT-DO NOT QUOTE OR CITE
-------
3000
2500-
2000-
1500-
1000-
500
PM
'2.5
geometric mean = 12.8 |jg/m3
an = 2.29
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Figure 3-13. Frequency distribution of 24-h average PM2 5 measurements obtained
from all California Air Resources Board dichotomous sampler sites
from 1989 to 1998.
3000
2500-
2000-
PM-lO-25
geometric mean = 15.7 (jg/m3
an = 2.26
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Figure 3-14. Frequency distribution of 24-h average PM10_2 5 concentrations
obtained from all California Air Resource Board Dichotomous
sampler sites from 1989 to 1998.
April 2002
3-25 DRAFT-DO NOT QUOTE OR CITE
-------
Riverside - Rubidoux
i uu -
140-
^ 120-
CO
E
"^ 100-
1 80-
03
-i— »
§ 60-
c
o
° 40-
20-
n
Fine
(n = 382)
ii i
y
T
1 r-
• .
y 1
\
r
Jan - Mar
IstQtr
Apr- Jun
2nd Qtr
Jul - Sept
3rd Qtr
Oct - Dec
4th Qtr
Figure 3-15. Concentrations of 24-h average PM2 5 measured at the Riverside-
Rubidoux site from 1989 to 1998. The data show the lowest, lowest
tenth percentile, lowest quartile, median (black squares), highest
quartile, highest tenth percentile, and highest PM2 5 values.
100
PM
2.5
geometric mean = 26.6 |jg/m3
a =2.10
VI
20 40
60
80
100
120 140 160 180
200
Figure 3-16. Frequency distribution of 24-h average PM25 concentrations
measured at the Riverside-Rubidoux site from 1989 to 1994.
April 2002
3-26
DRAFT-DO NOT QUOTE OR CITE
-------
1 considerations demonstrate the hazards in extrapolating conclusions about the nature of
2 variability in aerosol characteristics inferred at one location to another.
3
4 3.2.2 Diurnal (Orcadian) Variability in PM Concentrations
5 The variability of PM concentrations on time scales shorter than a day can, in principle, be
6 characterized by measurements made by continuous samplers (e.g, TEOMs and P-gauge
7 monitors that are currently used to provide Air Quality Index [AQI] information to the public).
8 A description of these methods was provided in Section 2.2.9. However, as shown in Chapter 2,
9 continuous methods are subject to artifacts because, in large part, of heating of their inlets to
10 remove water, which results in the loss of components such as ammonium nitrate and
11 semivolatile organic compounds (cf. Sections 2.2.2.1 and 2.2.3 for further details concerning the
12 chemistry of volatilizable components). Consequently, caution should be used in interpreting
13 results obtained by these techniques.
14 The composite diurnal variation of PM25 concentrations obtained throughout the
15 continental United States by 31 TEOM and P-gauge monitors reporting to AIRS in 1999 is
16 shown in Figure 3-17. As can be seen, there is a distinct pattern with maxima occurring during
17 the morning and evening. Notable exceptions to this pattern occur in California, where broad
18 nighttime maxima and daytime minima occur, which may be related to the use of p-gauges with
19 unheated inlets there. It should be noted in examining the diurnal variations shown in
20 Figure 3-17, that there is substantial day-to-day variability in the diurnal profile of PM25
21 measured at the same location that is smoothed out after a suitably long averaging period is
22 chosen. The large ratio of the interquartile range to the median values supports the view that
23 there is substantial variability in the diurnal profiles.
24 The diurnal variability of PM components is determined by interactions between variations
25 in emissions, the rates of photochemical transformations, and the vertical extent and intensity of
26 turbulent mixing near the surface. Wilson and Stockburger (1990) characterized the diurnal
27 variability of sulfate and lead in Philadelphia. At that time, Pb was emitted mainly by motor
28 vehicles. Pollutants emitted mainly by motor vehicles, such as carbon monoxide, show two
29 distinct peaks occurring during the morning and evening rush hours (see Chapter 3, U.S.
30 Environmental Protection Agency, 2000b). Pollutants, such as sulfate, which are transported
31 long distances in the free troposphere (i.e., above the planetary boundary layer), tend to be mixed
April 2002 3-27 DRAFT-DO NOT QUOTE OR CITE
-------
30-
E
D)
c
O
'ni 15'
i—
§
c
o
o
0-
«
03
+
:
6140 67m 6708 6709
-
-
' '
'""'
^~^
-
70
H-
E
7
-^
69
-
e
9
68
-
1 — 1
5 E
66
1 — 1
7 e
64
3
e
57
+
5
e
-
54
+
'— '
6576 6571 6605 E
—
-
-
~
-
6
640 6649
-
-
-
69
+
e
D
—
70
+
6 E
70
-
1 6
69
1
69
'
3
6704
l — l 6583
~ + ^ +
\ \ \ \ \ \ \ \ \ \
\\ \
\ \ \ \ \ \ \ \
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour
Figure 3-17. Intraday variability of hourly average PM2 5 concentrations across the
United States. Interquartile ranges, median and mean (+) values are
shown. Values above the box plots refer to the number of
observations during 1999.
Source: Fitz-Simons et al. (2000).
1
2
3
4
5
9
10
11
12
13
downward and have their highest concentrations during the afternoon when the intensity and
vertical extent of turbulent mixing (and chemical oxidation) are greatest. Secondary aerosol
components (such as secondary organic compounds) that are produced by photochemical
reactions may have a daily maximum in the afternoon, similar to ozone. PM produced by
residential heating (e.g., from wood burning), on the other hand, reach maximum levels during
the night when inversions are near the surface.
Although the interquartile ranges for hour-to-hour changes in PM2 5 concentrations shown
in Figure 3-17 encompass several Mg/m3, extreme values for the hour-to-hour variations can be
much larger (Fitz-Simons et al., 2000). The 98th percentile values for positive and negative
excursions in concentration are all less than 20 //g/m3. Maximum positive excursions were much
larger, ranging from 27 //g/m3 in the Northeast up to 220 //g/m3 in the Southwest and with
maximum excursions in other regions all less than 125 //g/m3. It should be borne in mind that
the hour-to-hour changes that are reported reflect the effects of a number of processes occurring
April 2002
3-28
DRAFT-DO NOT QUOTE OR CITE
-------
1 during passage through the sampler inlets and on the TEOM measurement elements. These
2 factors add uncertainty to the interpretation of the hour-to-hour changes that are observed, as
3 discussed in Chapter 2. However, because of the tendency of these monitoring instruments to
4 lose material by evaporation, the concentrations reported during excursions probably represent
5 lower limits to the true values that were present.
6
7 3.2.3 Relations Among Particulate Matter in Different Size Fractions
8 Relations Among PM25, PM10_2_S, and PM10
9 Data obtained in 1999 by collocated PM2 5 and PM10 FRM monitors have been used to
10 calculate the ratio of PM25 to PM10 concentrations and correlations among PM25, PM10_2 5, and
11 PM10 concentrations. Results are shown in Table 3-1 for each of the seven aerosol characteristic
12 regions identified in Chapter 6 of the 1996 PM AQCD. As can be seen from the table, the ratio
13 of PM25 to PM10 concentrations tends to be higher in the eastern United States than in the
14 western United States. This general pattern and the values are consistent with that found for the
15 studies included in Appendix 6A of 1996 PM AQCD. In that compilation based on the results of
16 studies using dichotomous samplers, the mean ratio of PM2 5 to PM10 was 0.75 in the East, 0.52
17 in the central United States, and 0.53 in the western United States. Although a large number of
18 paired entries have been included in Table 3-1, seasonal variations and annual averages in a
19 number of regions could not be determined from the data set because of data sparseness mainly
20 during the early part of 1999. It also can be seen in Table 3-1 that the ratio of PM2 5 to
21 PM10 was greater than one for a few hundred measurements. There are a number of reasons for
22 these results, as mentioned in Section 3.2.1 in the discussion on PM10_25 concentrations.
23
24 Ultrafine Particle Concentrations
25 Data for characterizing the concentrations of ultrafine particles (<0.10 //m Da) and the
26 relations between ultrafine particles and larger particles are sparse. Although ultrafine particles
27 dominate particle number concentrations, they make very minor contributions to PM2 5 mass.
28 For example, Cass et al. (2000) found that particles between 0.056 and 0.1 //m Da contributed
29 only 0.55 -1.16 //g/m3 at several sites in southern California. Perhaps the most extensive data
30 set for ultrafine particle properties is that described by Woo et al. (2001) for a site located 10 km
31 to the northwest of downtown Atlanta, GA. Size distributions from 3 to 2000 nm were measured
April 2002 3-29 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-1. DISTRIBUTION OF RATIOS OF PM25 TO PM10 AND CORRELATIONS BETWEEN PM25 AND PM10,
PM2 5 AND PM10 2 5, AND PM10 2 5 AND PM10 FOUND AT COLLOCATED MONITORING SITES IN SEVEN AEROSOL
to
o
o
to
OJ
OJ
o
O
H
6
o
0
H
O
Region
Northeast
Southeast
Industrial Midwest
Upper Midwest
Southwest
Northwest
Southern California
"Results considered to
Mean
0.70
0.70
0.70
0.53
0.38
0.50
0.47
Sites
45
76
92
39
23
73
36
384
CHARACTERISTIC (EPA/HEI) REGIONS IN 1999
Percentiles Correlations
Values 95 90 75 50 25 10 5 PM25:PM10 PM25:PM10_25
1433 0.97 0.95 0.77 0.67 0.60 0.51 0.48 0.72s 0.02
2823 1.27 1.06 0.74 0.63 0.54 0.46 0.43 0.69s -0.04s
4827 1.09 0.88 0.78 0.68 0.59 0.51 0.47 0.71s 0.17s
1446 0.92 0.84 0.62 0.49 0.44 0.34 0.24 0.35s -0.02
701 0.51 0.51 0.47 0.40 0.31 0.23 0.23 0.63s 0.49s
3300 0.67 0.65 0.56 0.49 0.44 0.39 0.36 0.69s 0.07s
1813 0.70 0.57 0.55 0.48 0.44 0.31 0.24 0.70s 0.19s
16,343
PM10.2.5:PM10
0.71s
0.69s
0.81s
0.93s
0.99s
0.77s
0.83s
be significantly different from zero at the a = 0.01 level.
Source: U.S. EPA Aerometric
Information Retrieval System.
O
HH
H
W
-------
1 every 12 minutes for 24 months beginning in August 1998. Approximately 89% of the total
2 number of particles were found to be smaller than 100 nm; whereas 26% were found to be
3 smaller than 10 nm. Concentrations tend to be lower during the summer than during the winter.
4 No correlation was found between number concentration and either volume or surface area for
5 particle sizes up to 2 //m. Because the total number of particles is concentrated in the smallest
6 size ranges, these results also indicate that fine particle mass does not correlate with the number
7 of ultrafine particles. The high time resolution of the measurements allows some inferences to be
8 made about the possible sources of the ultrafine particles. The number of particles larger than
9 10 nm tends to peak during the morning rush hour (around 8 a.m.) and then to decrease through
10 the day and to increase again after 6 p.m., consistent with a traffic-related source. Particles
11 smaller than 10 nm tend to peak during the mid-afternoon, consistent with nucleation involving
12 products of active photochemistry (McMurry et al., 2000). More direct relations between particle
13 mass observed in different size ranges can be obtained using multi-stage impactors. Keywood
14 et al. (1999) found a correlation between PM2 5 and PM015 of about 0.7; whereas they found
15 correlations of about 0.96 between PMl and PM2 5 and between PM2 5 and PM10 based on samples
16 collected by MOUDIs (Multiple Orifice Uniform Deposit Impactors) in six Australian cities.
17
18 3.2.4 Relations Between Mass and Chemical Component Concentrations
19 Time series of elemental composition data for PM2 5 based on X-ray fluorescence (XRF)
20 analyses have been obtained at a number of locations across the United States. Time series of
21 components of the organic carbon fraction of the aerosol have not yet been obtained. The results
22 of XRF analyses for the composition of the inorganic fraction of PM25 and PM10_2 5 are presented
23 in Table 3-2 for Philadelphia, PA and in Table 3-3 for Phoenix, AZ. The frequency distribution
24 for PM2 5 concentration data collected at these sites were shown in Figures 3-10 and 3-11.
25 All XRF analyses were performed at the same X-ray spectrometry facility operated by the U.S.
26 Environmental Protection Agency in Research Triangle Park, NC. Data shown in the first
27 column of Table 3-2 are based on analyses of filters collected over three years (April 1992 to
28 April 1995, labeled a) at the PBY site in southwestern Philadelphia. These data and data for
29 PM10 were collected using Harvard impactors. Data for PM2 5 and PM 10_2 5 shown in the second
30 and third columns were collected at the Castor Avenue Laboratory, operated by the City of
31 Philadelphia from July 25 to August 14, 1994, using a modified dichomotous sampler (VAPS).
April 2002 3-31 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-2. CONCENTRATIONS (in ng/m3) OF PM2 5, PM10 2 5, AND SELECTED ELEMENTS (ng/m3) IN THE
PM25 AND PM1025 SIZE RANGES WITH STANDARD DEVIATIONS (SD) AND CORRELATIONS BETWEEN
to
o
o
to
00
00
to
o
S
H
6
O
0
H
O
o
w
o
o
H
W
ELEMENTS AND PM2 5 MASS IN PHILADELPHIA, PA*
n =
1105
PM25>
Al
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Br
Pb
T~\oto /~\V\toi
Cone (ng/m3) ± SD
(unc)
17 ±0.9 (0.8) x 103
4.0 ±56 (31)
116 ±107 (21)
8.6 ± 14(10)
2100 ± 1610(143)
5.1 ±35 (3.4)
60.4 ± 45 (4.7)
47 ± 33 (4.2)
4.9 ±5.2 (4.1)
8.8 ±8.7 (1.8)
0.7 ±1.1 (0.7)
3.1 ±2.2 (0.8)
109 ±71 (10.5)
0.1 ±1.8 (1.4)
7.3 ±8.4 (1.4)
4.8 ±4.9 (1.1)
36.9 ±44 (3.7)
0.6 ±1.4 (1.2)
1.5 ±1.3 (0.6)
5.0 ± 11.7(0.9)
17.6 ±22 (2.5)
tr\f^f\ ot thi^ T5rv^oVv\rti^»i on n/~\
r
1
0.10
0.51
0.31
0.92
-0.01
0.50
0.39
0.44
0.37
0.15
0.39
0.50
0.04
0.22
0.25
0.21
0.18
0.63
0.11
0.19
me* in PniloH^
n = 20
PM252
Al
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Br
Pb
>lr\hio rVr\m AT
Cone (ng/m3) ± SD
(unc)
29.8 ± 14.7 (1.1) x 103
109 ±61 (21)
191 ± 134 (26)
15 ±4.3 (2.7)
3 190 ±1920 (207)
23 ± 28 (5.5)
68 ±21 (6.4)
63 ± 33 (9.0)
8.7 ±4.7 (9.0)
9.7 ±7.1 (2.9)
1.4 ± 1.2(2.9)
3.2 ±1.5 (1.6)
134 ±49 (0.5)
0.8 ±0.7 (8.5)
8.5 ±5.6 (0.3)
7.7 ±3. 8 (0.7)
56 ±37 (4.8)
0.4 ± 1.0(1.0)
1.3 ±0.8 (0.4)
14 ±12 (1.3)
28 ± 24 (2.4)
ii-il 1QQO tr, Ai-ii-il 1QCK TuitTi
r
1
0.15
0.22
0.72
0.91
0.19
0.31
-0.02
0.47
0.38
0.09
0.43
0.48
0.58
0.61
0.22
0.22
-0.02
0.65
0.21
0.26
n = 20
PM10.2,2
Al
Si
P
S
Cl
K
Ca
It
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Br
Pb
r\or-tr\rc
Cone (ng/m3) ± SD
(unc)
8.4 ±2.9 (0.4) x 103
325 ±241 (99)
933 ±652 (231)
28 ±9.4 (7.1)
38 ±45 (71)
47 ± 48 (5.8)
100 ±66 (10)
421 ±192 (31)
30 ±17 (5.6)
3.2 ±2.2 (1.5)
1.0 ±5.0 (0.9)
6.3 ±4. 1(0.6)
352 ±156 (24)
-0.2 ±0.5 (0.3)
2.0 ± 1.4(0.3)
14 ±12 (1.1)
52 ± 43 (4.7)
0± 0.5 (0.5)
-0.1 ±0.2 (0.2)
3.0 ±2.5 (0.5)
13 ±11 (1.3)
r
1
0.89
0.90
0.78
-0.15
-0.11
0.81
0.81
0.90
0.66
0.43
0.90
0.90
-0.10
0.08
-0.05
-0.03
0.07
-0.24
-0.10
0.10
I impactors.
2Data obtained at the Castor Avenue Laboratory, North Central Philadelphia from July 25 to August 14 with a modified dichotomous sampler.
'Note: Values in parentheses refer to analytical uncertainty (unc) in X-ray fluorescence determinations.
-------
TABLE 3-3. CONCENTRATIONS (in ng/m3) OF PM2 5, PM10 2 5 AND SELECTED
ELEMENTS IN THE PM2 5 AND PM10 2 5 SIZE RANGE WITH STANDARD
DEVIATIONS (SD) AND CORRELATIONS (r) BETWEEN ELEMENTS AND PM2 5
AND PM10 2 5 MASS IN PHOENIX, AZ*
n = 164
PM25
Al
Si
P
S
Cl
K
Ca
It
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Br
Pb
Cone (ng/m3) ± SD
(unc)
11.2 ±0.6 (0.6) x 103
125 ± 77 (30)
330 ±191 (48)
11 ±7.8 (5.7)
487 ± 254 (40)
19 ±44 (3.0)
110 ±63 (9.2)
129 ±72 (11)
11 ±7.1 (2.7)
0.7 ± 2.0 (2.2)
0.6 ± 0.9 (0.7)
5.7 ±4.3 (0.7)
177 ± 113(16)
-0.4 ±1.0 (1.0)
0.6 ±0.9 (0.5)
5.2 ±6.1 (1.5)
17 ±14.7 (1.8)
1.9 ±3.2 (0.6)
0.4 ± 0.8 (0.4)
3.8 ±2.0 (0.6)
6.6 ±6.6 (1.0)
r
1
0.23
0.35
0.52
0.16
0.13
0.67
0.51
0.44
-0.28
0.41
0.64
0.80
-0.01
0.38
0.69
0.64
0.50
0.40
0.57
0.69
n = 164
PM10.2.5
Al
Si
P
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zn
As
Se
Br
Pb
Cone (ng/m3) ± SD
(unc)
27.6 ± 14.8 x 103
1879 ± 979 (547)
535 ±2825 (1347)
37 ±20 (17)
131 ±47 (26)
208 ± 204 (24)
561 ±298 (62)
1407 ±755 (124)
130 ±71 (20)
2.0 ±2.0 (1.5)
2.6 ±1.7 (0.7)
29 ± 16(3.0)
1211 ±674 (133)
1.2 ±2.2 (1.9)
1.8 ±1.4 (0.7)
10.3 ±9.0 (1.5)
25 ± 16(3.2)
0.6 ±0.8 (0.6)
-0.02 ±0.3 (0.3)
0.8 ± 0.6 (0.4)
4.6 ±3. 8 (1.1)
r
1
0.92
0.92
0.58
0.77
0.28
0.92
0.90
0.90
0.51
0.76
0.91
0.90
0.38
0.70
0.58
0.64
0.41
0.21
0.48
0.59
* Values in parenthesis refer to analytical uncertainty (unc) in X-ray fluorescence determinations.
Source: Zweidingeretal. (1998).
1 The samples at the Phoenix site were collected in 1996 and 1997 using the same type of
2 dichotomous sampler used in the shorter term study in Philadelphia. These data are shown to
3 give an idea of the range of concentrations found in studies conducted more recently than those
April 2002
3-33
DRAFT-DO NOT QUOTE OR CITE
-------
1 shown in Appendix 6A of the 1996 PM AQCD. The speciation network will at least provide
2 more thorough coverage of the composition of particles in the PM25 size range across the United
3 States. Results from the pilot study for the speciation network are given in Appendix 3B.
4 As can be seen from inspection of Tables 3-2 and 3-3, the analytical uncertainty (given in
5 parentheses next to concentrations) as a fraction of the absolute concentration is highly variable.
6 It exceeds the concentration for a number of trace metals whose absolute concentrations are low;
7 whereas it is very small for abundant elements such as sulfur.
8 Sulfur is the major element analyzed in the PM25 size fraction in the two Philadelphia
9 studies and is highly correlated with PM2 5; however its abundance is roughly two orders of
10 magnitude lower in the PM10_2 5 size range and is negatively correlated with PM10_2 5.
11 Concentrations of the crustal elements: Al, Si, K, Ca, and Fe are much higher in the PM10_25 size
12 range than in the PM25 size range and are well correlated with PM10_25. A number of trace
13 elements (e.g., Cr, Co, Ni, Cu, Zn, As, Se and Pb) are detectable in the two PM2 5 data sets, and
14 the concentrations of many of these elements are much greater than the uncertainty in their
15 determination. Except for Co, As, and Se which are not detected in the PM10_25 samples, the
16 concentrations of many elements (Cr, Zn, and Pb) are comparable in the PM25 and PM10_25 size
17 ranges. The concentration of Cu is significantly higher in the PM10_2 5 size range, whereas the
18 concentration of Ni is smaller in the PM10_25 size range than in the PM25 size range.
19 There are a number of distinct differences between the PM2 5 sets for Philadelphia and
20 Phoenix. For instance, sulfate and associated cations and water that would be expected to
21 correspond to the measurement of S appear to constitute a major fraction of the composition of
22 the PM in the Philadelphia data set; whereas they appear to constitute a much smaller fraction of
23 the PM in the Phoenix data set. The highest PM2 5 values were observed in Philadelphia during
24 episodes driven by high sulfate abundances; whereas those in Phoenix were driven by raised soil
25 dust. The concentration of S in Phoenix is much lower in the Phoenix PM2 5 data set than in
26 either Philadelphia PM2 5 data set, and it is only weakly correlated with PM2 5. As in
27 Philadelphia, the concentration of S in Phoenix is higher in the PM2 5 size range than in the
28 PM10_2 5 size range. Trace metals (e.g., Cr, Co, Ni, Cu, Zn, As, and Pb) are not well correlated
29 (0.04 < r < 0.25) with PM2 5 in the Philadelphia data set; whereas they are more variably
30 correlated (0.01 < r < 0.69) with PM25 in the Phoenix data set. The uncertainty in the
31 concentration measurement most probably plays a role in determining a species' correlation with
April 2002 3-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 PM2 5, especially when the analytical uncertainty is high relative to concentration, as it is for a
2 number of elements in the data shown in Tables 3-2 and 3-3. Concentrations of Al, Si, K, Ca,
3 and Fe are again much higher in the PM10_2 5 size range than in the PM2 5 size range and are
4 strongly correlated with PM10_2 5 in both data sets.
5 There are also similarities in the PM2 5 data sets for Philadelphia and Phoenix. Crustal
6 elements are not as well correlated with PM2 5 as they are with PM10_2 5 in both data sets. The
7 concentrations of trace metals (Cr, Ni, Cu, and Zn) in PM25 are similar in Philadelphia and
8 Phoenix. It can also be seen that their concentrations are of the same order of magnitude in both
9 PM2 5 and PM10_2 5. Concentrations of Cu are noticeably higher in PM10_2 5 than in PM2 5 in both
10 Philadelphia and Phoenix. These results are consistent with those of many monitoring studies
11 shown in Appendix 6A of the 1996 PM AQCD, which also show that concentrations of these
12 metals are of the same order of magnitude in both size fractions and that concentrations of Cu
13 tend to be higher in PM10_25 than in PM2 5.
14 One study suggests that the partitioning of trace metals between the fine and coarse
15 fractions varies with PM concentration. Salma et al. (2002) determined the size distribution of a
16 number of trace elements at four sites characterizing environments ranging from the urban
17 background to an urban traffic tunnel in Budapest, Hungary. S, K, V, Ni, Cu, Zn, As, and Pb
18 were found mainly in the fine fraction at the urban background site; but their mass median
19 aerodynamic diameters increased with increasing PM concentrations until they were all found
20 mainly in the coarse fraction in the traffic tunnel. They also found that Na, Mg, Al, Si, P, Ca, It,
21 Fe, Ga, Sr, Zr, Mo, and Ba were concentrated mainly in the coarse fraction at all four sites and
22 that their mass median aerodynamic diameters increased with increasing PM concentrations.
23 The mean concentration of Pb observed in the methods evaluation study for the speciation
24 network was only about 5 ng/m3 in Philadelphia during the first half of 2000 (cf. Appendix 3B);
25 whereas its concentration was about three times higher during the studies conducted during the
26 early 1990s (Table 3-3). In a study conducted in the greater Philadelphia area during the summer
27 of 1982, Dzubay et al. (1988) found concentrations of Pb of about 250 ng/m3, or about fifty times
28 higher than observed in 2000. The mean Pb concentration was about 3 ng/m3 at the Phoenix site
29 included as part of the same methods evaluation study for the speciation network; however, the
30 mean Pb concentration was 39 ng/m3 during an earlier study conducted during 1989 and 1990 in
31 Phoenix (Chow et al., 1991). These changes in Pb concentrations are consistent with those in
April 2002 3-35 DRAFT-DO NOT QUOTE OR CITE
-------
1 many other urban areas for which monitoring studies have been conducted during the late 1970s
2 and 1980s (cf. Appendix 6 A of the 1996 PM AQCD) and for which there are data given in
3 Appendix 3B. It should be remembered that the older studies were conducted while Pb was still
4 used as a gasoline additive. The ratio of Pb in PM2 5 to Pb in PM10_25 was also much higher in the
5 older studies than in the more recent ones, reflecting the importance of combustion as its source.
6 Smaller decreases are apparent in the concentrations of other trace metals such as Cu, Ni, and Zn
7 between studies conducted in the early 1980s and in the methods evaluation study for the
8 speciation network conducted in 2000.
9 Some indication of the sources of metals such as Pb, Cu, Cd, and Zn in current, ambient
10 PM25 and PM10_2 5 samples can be obtained by examining their sources in urban runoff. The
11 sources of these elements in urban runoff were found to be the weathering of building surfaces,
12 motor vehicle brake and tire wear, engine oil and lubricant leakage and combustion, and wet and
13 dry atmospheric deposition (Davis et al., 2001). Once deposited on the ground, these elements
14 can be resuspended with other material as PM2 5 and PM10_25, although research is needed into the
15 mechanisms of how this is accomplished. Wind-abrasion on building siding and roofs (coatings
16 such as Pb paint and building material such as brick, metal, and wood siding); brake wear (brake
17 pads contain significant quantities of Cu and Zn); tire wear (Zn is used as a filler in tire
18 production); and burning engine oil could all produce particles containing these metals,
19 especially Zn.
20 Data for the chemical composition of ambient ultrafine particles are sparse. In a study
21 conducted at several urban sites in Southern California, Cass et al. (2000) found that the
22 composition of ultrafine particles ranged from 32 to 67% organic compounds, 3.5 to 17.5%
23 elemental carbon, 1 to 18% sulfate, 0 to 19% nitrate, 0 to 9% ammonium, 1 to 26% metal oxides,
24 0 to 2% sodium, and 0 to 2% chloride. Thus carbon, in various forms, was found to be the major
25 contributor to the mass of ultrafine particles. However, ammonium was found to contribute 33%
26 of the mass of ultrafine particles at one site in Riverside. Iron was the most abundant metal
27 found in the ultrafine particles. Chung et al. (2001) found that carbon was the major component
28 of the mass of ultrafine particles in a study conducted during January of 1999 in Bakersfield, CA.
29 However, in the study of Chung et al., the contribution of carbonaceous species (OC and EC)
30 (typically 20 to 30%) was much lower than that found in the cities in Southern California. They
31 found that calcium was the dominant cation, accounting for about 20% of the mass of ultrafine
April 2002 3-36 DRAFT-DO NOT QUOTE OR CITE
-------
1 particles in their samples. Sizable contributions from silicon (0 to 4%) and aluminum (6 to 14%)
2 were also found. Further studies, including scanning electron microscopy, may be needed to
3 quantify the role of coarse particle bounce from the upper stages of their MOUDI impactor.
4 Gone et al. (2000) measured the size distribution of trace elements from 0.056 //m to
5 1.8 //m Da in Pasadena, CA and in the Great Smoky Mountains National Park, TN. They found
6 that elements identified as being of anthropogenic origin had mass median diameters below 1 //m
7 PM; whereas elements of crustal origin generally had a mass median diameter greater than 1 //m.
8 Concentrations of trace metals were much higher in the accumulation mode than in the ultrafine
9 mode in both study areas. In PMl3 76% of Cr, 95% of Fe, 94% of Zn, 89% of As, and 79% of Cd
10 at the Tennessee site were found in the accumulation mode; and 70% of Fe, 85% of Zn, 92% of
11 As, and 84% of Cd were found in the accumulation mode in Pasadena. Fe was the most
12 abundant metal found in the ultrafine particles. The abundance of crustal elements, such as Al,
13 declined rapidly with decreasing particle size at both locations, and Al in PMX probably
14 represented the lower tail of the coarse PM mode. However, on two days at Pasadena there were
15 increases in the concentration of Al in ultrafine particles that were associated with increases in Sc
16 and Sm. The latter two elements originate exclusively from crustal material (Gone et al., 2000).
17
18 3.2.5 Spatial Variability in Particulate Matter and its Components
19 PM25
20 Aspects of the spatial variability of PM2 5 concentrations on the urban scale are examined in
21 this section. Intersite correlation coefficients for PM25 can be calculated based on the results of
22 FRM monitors placed at multiple sites within Metropolitan Statistical Areas (MSAs) across the
23 United States. Pearson correlation coefficients (r) calculated for pairs of monitoring sites in the
24 Columbia, SC; Detroit, MI; Chicago, IL; and Los Angeles, CA MSAs are shown in Table 3-4.
25 The 90th percentile value, P90, of the absolute differences (in //g/m3) between the two sites is
26 shown below r along with the coefficient of divergence (COD) in parentheses, and the number of
27 observations used in the calculation of r, P90 and COD is given on the third line. The COD was
28 used by Wongphatarakul et al. (1998) as a measure of the degree of similarity between two
29
30
April 2002 3-37 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-4. MEASURES OF THE SPATIAL VARIABILITY OF PM2 5
CONCENTRATIONS WITHIN SELECTED METROPOLITAN
STATISTICAL AREAS
(a) Columbia, SC
Site ID. #
45-063-0005
45-063-0008
45-079-0007
45-079-0019
Mean
Obs
SD
(b) Detroit, MI
Site ID. #
26-099-0009
26-125-0001
26-147-0005
26-163-0033
26-163-0036
Mean
Obs
SD
45-063-0005 45-063-0008
1 0.882
(5.3,0.121)
215
1
Key
AIRS Site I.DJ
Pearson r
45-079-0007
0.949
(3.9,0.081)
204
0.933
(4.0, 0.082)
202
(90th %-ile difference in concentration, coefficient of divergence)
number of observations
14.680 16.462
231 228
6.760 7.121
26-099-0009 26-125-0001
1 0.958
(4.9, 0.107)
83
1
Key
AIRS Site I.D.#
Pearson r
15.461
216
6.900
26-147-0005
0.952
(5.6,0.127)
96
0.939
(5.8,0.121)
73
1
(90th %-ile difference in concentration, coefficient of divergence)
number of observations
13.450 15.552
113 90
7.922 9.223
14.172
102
8.771
45-079-0019
0.93
(4.8, 0.099)
216
0.949
(3.3, 0.067)
216
0.971
(2.7, 0.06)
203
1
16.098
229
7.148
26-163-0033
0.931
(12.7, 0.222)
98
0.92
(12.3,0.193)
77
0.876
(13.3, 0.222)
89
1
20.173
108
10.475
26-163-0036
0.926
(9.0, 0.177)
96
0.917
(8.3,0.151)
75
0.875
(8.9,0.197)
88
0.923
(7.1,0.108)
89
1
17.446
103
9.626
April 2002
3-38
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-4 (cont'd). MEASURES OF THE SPATIAL VARIABILITY OF PM2 s CONCENTRATIONS WITHIN
SELECTED METROPOLITAN STATISTICAL AREAS
to
o
o
to
OJ
VO
fe
H
6
O
o
H
O
o
H
W
O
O
(c) Chicago, IL
Site ID. # 17-031-0014 17-031-0022 17-031-0050 17-031-0052
17-031-0014 1
17-031-0022
17-031-0050
17-031-0052
17-031-1016
17-031-2001
17-031-3301
17-031-4006
17-031-4201
17-043-4002
17-197-1002
Mean 15.823
Obs 104
SD 7.935
0.912 0.946 0.909
(4.4,0.121) (4.6,0.077) (6.6,0.13)
96 78 100
1 0.92 0.872
(5.4,0.113) (6.5,0.14)
87 108
1 0.941
(5.0, 0.094)
259
1
Key
AIRS Site I.DJ
Pearson r
17-031-1016
0.921
(7.5,0.143)
92
0.866
(7.0, 0.141)
103
0.93
(7.8,0.12)
83
0.887
(7.9, 0.133)
105
1
(90th %-ile difference in concentration, coefficient of divergence)
number of observations
17.933 16.996 18.295
113 274 346
8.175 8.468 9.289
20.277
108
9.331
17-031-2001
0.902
(5.6,0.111)
98
0.892
(5.7,0.131)
104
0.955
(3.5, 0.082)
89
0.885
(7.3,0.125)
109
0.932
(7.3,0.108)
99
1
16.790
113
7.694
17-031-3301
0.927
(5.1,0.104)
98
0.879
(6.0, 0.132)
106
0.923
(5.3, 0.096)
91
0.881
(7.0,0.128)
110
0.898
(7.5,0.124)
102
0.931
(4.5, 0.084)
110
1
16.889
115
7.689
17-031-4006
0.876
(5.8,0.133)
88
0.689
(7.9, 0.213)
92
0.75
(7.9, 0.176)
75
0.797
(8.5,0.177)
98
0.787
(10.0, 0.205)
92
0.861
(5.9,0.153)
93
0.823
(7.0,0.158)
95
1
15.268
101
8.423
17-031-4201
0.936
(5.3,0.139)
95
0.86
(7.9,0.197)
101
0.928
(6.2,0.162)
247
0.879
(9.6,0.179)
310
0.915
(9.8, 0.2)
98
0.943
(5.5,0.14)
101
0.915
(6.4, 0.152)
103
0.818
(7.3, 0.146)
92
1
14.283
327
7.905
17-043-4002
0.885
(5.7,0.13)
95
0.855
(7.2, 0.165)
100
0.922
(5.3,0.117)
91
0.836
(8.5,0.154)
112
0.902
(9.5,0.154)
95
0.949
(4.3,0.1)
99
0.953
(4.4, 0.092)
101
0.865
(5.1,0.124)
88
0.922
(4.8,0.123)
106
1
15.215
116
7.568
17-197-1002
0.774
(7.4,0.158)
81
0.79
(7.1,0.17)
87
0.867
(7.6,0.131)
87
0.721
(10.2,0.169)
108
0.84
(10.5,0.173)
85
0.893
(5.1,0.118)
89
0.873
(5.8,0.128)
91
0.752
(7.6, 0.161)
78
0.809
(7.1,0.157)
99
0.921
(4.2, 0.099)
90
1
15.994
112
7.405
-------
1
2
3
TABLE 3-4 (cont'd). MEASURES OF THE SPATIAL VARIABILITY OF PM2 5
CONCENTRATIONS WITHIN SELECTED METROPOLITAN
STATISTICAL AREAS
(d) Los Angeles, CA
Site ID. # 06-037-0002 06-037-1103 06-037-1601 06-037-4002 06-037-9002
06-037-0002 1 0.828 0.763 0.573 0.276
(12.8,
0.192) (17.3,0.211) (20.2,0.263) (28.0,0.392)
391 196 379 186
06-037-1103
06-037-1601
K
AIRSS
Pear
06-037-4002 (90«, 0/0_ile difference in concent
number of c
06-037-9002
Mean 21.682 22
1 0.88 0.752 0.328
(11.8,0.140) (14.6,0.191) (26.4,0.375)
173 353 164
1 0.859 0.363
11.8,0.174 31.0,0.411
171 181
ey
te I.D.#
sonr
-ation, coefficient of divergence) * 0.338
bservations C24 4 03561
157
1
207 24.764 20.225 10.917
Obs 469 428 218 417 204
SD 13.923 13.
840 14.056 12.994 5.043
Source: Pinto et al., (2002). Data from U.S. EPA Aerometric Information Retrieval System (AIRS).
aerosol data sets1. The annual mean concentrations, the number of observations used to calculate
the annual average, and the standard deviation are shown directly beneath the correlation tables
for each site. These analyses along with those for another 23 MSAs are given along with maps in
'The COD for this purpose is defined as follows:
(3-1)
where x^ and x4 represent the 24-h average PM2 s concentration for day i at site j and site k and p is the number of
observations.
April 2002
3-40
DRAFT-DO NOT QUOTE OR CITE
-------
1 Appendix 3A. The four MSAs shown in Table 3-4 were chosen to illustrate different patterns of
2 spatial variability across the United States. In addition, air pollution-health outcome studies have
3 been performed in a few of these MSAs. It can be seen from inspection of Table 3-4 that
4 correlation coefficients vary over a wide range in the MSAs shown. Correlations between sites
5 in the Columbia, SC MSA and the Detroit, MI MSA are all high and span a relatively narrow
6 range (0.88 to 0.97). Correlations between sites in the Chicago, IL MSA span a wider range
7 (0.69 to 0.96). However, the correlations between sites in the Los Angeles-Long Beach MSA are
8 much lower than in the three other MSAs and span an even wider range of values (0.28 to 0.88).
9 The extension of these analyses to include the relevant CMS As (consolidated MSA) would also
10 produce a number of sites that are even less well correlated with each other in part because of the
11 larger distances involved. Correlation coefficients between pairs of sites in the other 23 MSAs
12 given in Appendix 3A fall within the range of values given in Table 3-4. Some indication of the
13 performance of collocated monitors is given by inspection of the last two columns of
14 Table 3A-10. These data were obtained by two collocated PM2 5 monitors in the Steubenville,
15 OH-Weirton, WV MSA. Values of r, P90, and COD for these two monitors are 0.978, 2.5 //g/m3,
16 and 0.101.
17 There may be a regional pattern evident in the data given in Appendix 3 A, data for which
18 correlations tend to be higher between monitoring sites in MSAs in the eastern and central
19 United States than between monitoring sites in the western United States. In a few MSAs
20 (Milwaukee, WI; Norfolk, VA; Grand Rapids, MI; and Baton Rouge, LA), intersite correlations
21 are all greater than 0.9. In several others (Philadelphia, PA; Columbia, SC; Steubenville, OH;
22 Detroit, MI; Kansas City, KS-MO; and Dallas, TX), they are all greater than 0.8. Intersite
23 correlations tend to be lower and to span a broader range in several cities such as Atlanta, GA;
24 Seattle, WA; and Los Angeles, CA, in part due to the location of monitoring sites outside of the
25 main urban area and in a different air shed. In many MSAs, there is a wide range in the intersite
26 correlations that are found. For example, in the Seattle, WA MSA (Table 3 A-23), values r of
27 range from 0.41 to 0.95. Correlations between sites in the Atlanta, GA, Birmingham, AL, and
28 Tampa, FL MSAs tend to be lower and span a broader range than do those for the other southern
29 cities examined (Columbia, SC; Norfolk, VA; Baton Rouge, LA; and Dallas, TX). Likewise,
30 correlations between a number of sites in western MSAs are higher than those in some eastern
April 2002 3-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 MSAs. For example, correlations between monitors in the Pittsburgh, PA MSA tend to be lower
2 than those in the Salt Lake City, UT MSA.
3 There are a number of factors that affect intersite correlations within MSAs. These include
4 field measurement and laboratory analysis errors, placement of monitors close to active sources,
5 placement of monitors in outlying areas, placement of monitors in locations that are isolated
6 topographically from other monitors, placement of monitors in areas outside of local atmospheric
7 circulation regimes (e.g., land-sea breezes), and transient local events (thunderstorms, sporadic
8 emissions).
9 It should not be automatically assumed that distance between sites in urban areas is solely
10 responsible for the spatial variability that is observed. In several areas such as Atlanta, GA;
11 Seattle, WA; and Los Angeles-Long Beach, CA, there is at least one site that is remote from the
12 others (by at least 100 km) and is physically separated from them by mountains and is really not
13 part of the urban area. Correlations between concentrations at these sites and others tend to be
14 lower than among the other sites, and concentration differences tend to be larger. However, in
15 many MSAs, especially in the East, correlations are higher, and differences in concentrations are
16 lower for sites that are located farthest apart. This situation arises because these sites are
17 influenced more by the regional background of secondary PM rather than by local sources. Nor
18 is there any set distance below which correlations and differences in concentrations tend towards
19 some limiting values. In Gary, IN, for example, intersite correlations are lowest, and
20 concentration differences are highest for the closest site pair.
21 Indications of land use (commercial, industrial, residential, agricultural, forest) and location
22 of sites (urban/city center, suburban, rural) are given in the AIRS data base. Categories such as
23 urban/city center can refer to very different conditions in Columbia, SC and Chicago, IL. Also, it
24 should not be automatically assumed that concentrations measured at sites categorized as
25 industrial are dominated by local emissions. The PM2 5 monitoring sites are generally deployed
26 to capture potential population exposures in a variety of environments as opposed to monitoring
27 for compliance as it exists around local sources. It should be remembered that much of PM25 is
28 secondary in origin. The widespread formation of secondary PM coupled with the long lifetime
29 of PM25 ensures some measure of uniformity in the correlations of PM25 across urban areas.
30 Correlations between many site pairs classified as industrial can be high even though they are
31 separated by large distances, as in the Seattle MSA.
April 2002 3-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 Some indication of the variability of primary PM25 produced by local sources can be
2 obtained by examining the variability of carbon monoxide (CO), which is produced mainly by
3 mobile sources (U.S. Environmental Protection Agency, 2000b) and by the variability in
4 elemental carbon (EC) concentrations (Kinney et al., 2000). CO is relatively inert on the urban
5 scale, and its distribution is governed by the spatial pattern of its emissions and the subsequent
6 dispersion of these emissions and not by photochemistry. Carbon monoxide concentrations are
7 at least a factor of three higher near urban centers than in surrounding rural areas within the four
8 consolidated metropolitan statistical areas examined in the EPA document, Air Quality Criteria
9 for Carbon Monoxide (CO AQCD) (U.S. Environmental Protection Agency, 2000b).
10 Correlations of CO within the urban areas examined in that document were all low to moderate.
11 Therefore, it might be expected that primary PM2 5 produced by local traffic should be at least as
12 heterogeneous as CO in a given urban area. EC is a significant component of diesel exhaust (cf.
13 Appendix 3D). Kinney et al. (2000) measured EC and PM2 5 concentrations at four sites located
14 on sidewalks of streets characterized by varying exposures to diesel emissions in upper
15 Manhattan (Harlem, NY). Whereas the mean PM25 concentrations varied by about one-third
16 from 37 to 47 //g/m3 at the four sites, mean EC concentrations varied by a factor of four from
17 1.5 to 6.2//g/m3. The corresponding ratios of EC to PM25 ranged from 0.039 to 0.14. Although
18 EC constituted a relatively small fraction of PM25 in this study, spatial variability in its sources
19 (diesel and gasoline fueled vehicles, resuspended road dust, and cooking) contributed, on
20 average, about one-third of the spatial variability observed in PM25 concentrations. Further
21 analyses are needed to determine whether the remaining variability could be attributed to other
22 local and city-wide sources. Because the effects of emissions from local point sources on
23 receptor sites depend strongly on wind direction, correlations involving contributions from
24 sources can be much lower than from area sources (much as motor vehicle traffic) or from
25 regionally dispersed sources (such as the photochemical production of secondary organic PM and
26 sulfate).
27 The difference in mean PM2 5 concentrations between the site with the lowest and the site
28 with the highest mean concentration range in all MSAs included in Appendix 3A ranges from
29 less than 1 //g/m3 to about 7 //g/m3, except for the Los Angeles MSA which shows larger
30 differences. In the Los Angeles MSA, there is one monitoring site (Figure 3A-25a) that is
31 separated from the remaining sites by the San Gabriel Mountains and has much lower mean
April 2002 3-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 PM2 5 concentrations, much smaller seasonal variability in concentrations, and much lower
2 maximum concentrations than these other sites. However, the annual mean concentrations at all
3 the other sites within the Los Angeles MSA are within 5 //g/m3 of each other. Differences in
4 annual mean concentrations are also larger between sites located in different MSAs but within
5 the same CMS A. For example, in the consolidated MSA of Los Angeles-Riverside the range of
6 annual mean PM2 5 concentrations is extended from about 20 //g/m3 in the urban area of
7 Los Angeles county to about 29 //g/m3 in Riverside County. Large differences in annual mean
8 concentrations within a given area reflect differences in source or meteorological or unique
9 topographic characteristics affecting sites; whereas very small differences found in some areas
10 may only be the result of measurement imprecision.
11 Whereas high correlations of PM2 5 provide an indication of the spatial uniformity in
12 temporal variability (directions of changes) in PM2 5 concentrations across urban areas, they do
13 not imply uniformity in the PM2 5 concentrations themselves. The 90th percentile difference in
14 concentrations (P90) and the coefficient of divergence are used here to give a more quantitative
15 indication of the degree of spatial uniformity in PM25 concentrations across urban areas. A COD
16 of zero implies that both data sets are identical, and a COD of one indicates that two data sets are
17 completely different. The calculation of the Pearson correlation coefficient, P90, and COD allows
18 for distinctions between pairs of sites to be made based on various combinations of these
19 parameters. Figure 3-18 shows examples of the varying degree of heterogeneity in
20 concentrations between pairs of sites that are highly correlated (r > 0.9 for all three site pairs).
21 The increase in the spread of concentrations between the chosen site-pairs is reflected in
22 increases in both P90 and COD. Pairs of sites showing low correlations, values of P90 > 10 //g/m3,
23 and CODs > 0.2, as in Los Angeles, CA (Table 3-5), indicate heterogeneity in both PM25
24 concentrations and in their temporal variations. Note that the extended urban area or the CMSA
25 includes Riverside County, as well as Los Angeles County. Even lower correlations and a
26 greater degree of heterogeneity in PM25 concentrations were found in the extended CMSA. Pairs
27 of sites showing high correlations and CODs < 0.1 and P90's < 5 //g/m3 (as in Columbia, SC)
28 indicate homogeneity in both PM2 5 concentrations and in their temporal variations. Presumably,
29 sites such as these are more strongly affected by regional than to local sources. Pairs of sites
30 showing high correlations (r > 0.9) and CODs > 0.2 and P90's > 10 //g/m3 (as in Detroit, MI)
31 indicate heterogeneity in concentrations but homogeneity in their day to day changes.
April 2002 3-44 DRAFT-DO NOT QUOTE OR CITE
-------
Columbia SC 1999 & 2000
03
O
1
O
o
O
"5
r
V
Chicago IL2000
$
c
fc
o
o
O
"S
E
Z
30 -,
25 -
20 -
15 -
10 -
5 -
n
17-031-2001 vs 17-031 -420
r = 0.94
COD = 0.14
Pso = S.Sftg/nf
I.M -
Detroit Ml 2000
| 12 -
£ 10-
§ 8-
O
* 6-
oi 4 -
1 2-
z n
I
26-099-0009 vs 26-163-0033
r = 0.93
COD = 0.22
P9D= 12.7/ig/m3
||
I. .1 . 1
Concentration Difference (/ig/m3)
Figure 3-18. Occurrence of differences between pairs of sites in three MSAs. The absolute
differences in daily average PM2 5 concentrations between sites are shown on
the x-axis and the number of occurrences on the y-axis. The MSA, years of
observations, AIRS site LD. numbers for the site pairs, Pearson correlation
coefficients (r), coefficients of divergence (COD), 90th percentile (P90)
difference in concentration between concurrent measurements are also
shown.
Source: Pinto et al. (2002)
April 2002
3-45
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-5. MEASURES OF THE SPATIAL VARIABILITY OF PM10 2 5
CONCENTRATIONS WITHIN SELECTED METROPOLITAN
STATISTICAL AREAS
(a) Detroit, MI
Name
26-163-0001
26-163-0015
26-163-0025
Mean
Obs
SD
26-163-0001 26-163-0015 26-163-0025
1 0.576 0.542
53 50
1 0.393
51
1
11.517 19.416 7.328
56 58 55
10.262 15.611 7.638
(b) Chicago, IL
Name
17-031-1016
17-031-2001
17-031-3301
17-197-1002
Mean
Obs
SD
17-031-1016 17-031-2001 17-031-3301 17-197-1002
1 0.69 0.544 0.583
49 51 43
1 0.865 0.823
54 44
1 0.777
46
1
16.259 14.475 17.812 6.894
93 56 58 49
18.972 12.137 13.641 10.217
(c) Los Angeles
Name
06-037-1002
06-037-1103
06-037-4002
06-037-9002
Mean
Obs
SD
06-037-1002 06-037-1103 06-037-4002 06-037-9002
1 0.79 0.83 0.59
51 49 43
1 0.79 .042
53 46
1 0.39
47
1
19.1 20.3 19.3 15.6
52 55 56 52
10.58 8.4 9.2 12.9
April 2002
3-46
DRAFT-DO NOT QUOTE OR CITE
-------
1 Conversely, in the Tampa, FL MSA pairs of sites are only moderately correlated (0.6 < r < 0.7),
2 but the distribution of concentrations is rather homogeneous (COD < 0.1) (cf Appendix 3 A).
3 Thus, a number of different combinations of spatial uniformity in PM2 5 concentrations and
4 correlations of these concentrations are found.
5 Values of P90 for absolute differences in concentrations between sites span a wide range in
6 the data set given in Appendix 3 A. In many instances they can be quite low, only about a few
7 Mg/m3; these cases are found mainly in the eastern United States. Values of P90 can be greater
8 than 40 //g/m3; these cases are found mainly in the western United States. Maximum differences
9 in concentrations between sites can be much larger than shown in Figure 3-18 and have been
10 larger than 100 //g/m3 on several occasions in the Atlanta, GA and Los Angeles-Long Beach, CA
11 MSAs. Rizzo and Pinto (2001) and Fitz-Simons et al. (2000) examined correlations between
12 sites located even farther apart than those examined here based on the 1999 AIRS data set for
13 PM25. They found that in a number of MSAs, PM2 5 concentrations are still well correlated
14 (r >0.7) to distances of 100 km or more. Leaderer et al. (1999) found r = 0.49 between sites
15 outside of homes and a regional background monitor located from 1 to 175 km away in
16 southwestern Virginia. PM2 5 tends to be correlated over much larger areas in the East than in the
17 West, mainly because the terrain tends to be flatter over wider areas in the East (Rizzo and Pinto,
18 2001).
19 There is also evidence for inter-annual variability in the spatial variability in PM2 5
20 concentrations. The median year-to-year changes in inter-site r (0.03), P90 (-0.75 //g/m3), and
21 COD (-0.015) from 1999 to 2000 do not differ significantly from zero for all the site pairs
22 considered in Appendix 3 A. The year-to-year changes in the spatial variability of PM25
23 concentrations in a number of MSAs such as the Columbia, SC; Grand Rapids, MI; Milwaukee,
24 WI; Baton Rouge, LA; Kansas City, KS-MO; Boise, ID; and Portland, OR MSAs are similar and
25 are smaller than those found in the Cleveland, OH; Salt Lake City, UT; and San Diego, CA
26 MSAs. The ranges in these parameters are largest for a number of individual site-pairs,
27 especially those involving sites that are remote from the others in their MSAs. In these MSAs
28 (such as the Atlanta, GA; Los Angeles, CA; and Seattle, WA MSAs) there are sites that may be
29 located in different air sheds from the remaining sites. Year-to-year changes in parameters
30 describing spatial variability in PM2 5 concentrations tend to be larger when sites in different
31 counties within a given MSA are considered rather than when sites in the same county are
April 2002 3-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 considered. There are a number of factors that can account for inter-annual variability in these
2 parameters such as changes in patterns in the emissions of primary PM25, in the transport and
3 rates of transformation of secondary PM2 5 precursors in field measurement and analysis
4 procedures.
5 Some additional data for indicating the stability with respect to year to year changes in
6 spatial variability are available from earlier studies. For example, a comparison of data obtained
7 during the summers of 1992 and 1993 (Wilson and Suh, 1997) as shown in Figure 3-19 and data
8 obtained during the summer of 1994 (Pinto et al., 1995) (cf. Table 3-8) in Philadelphia, PA
9 suggests that inter-site correlations of PM2 5 have remained high and that they changed very little
10 between the two study periods.
11
12 PM^S
13 Intersite correlations of PM10_2 5 concentrations obtained during the summers of 1992 and
14 1993 in Philadelphia, PA (Wilson and Suh, 1997) are shown in Figure 3-19. As can be seen,
15 correlations of PM10_2 5 are substantially lower than those for PM2 5.
16 Intersite correlation coefficients can also be calculated for PM10_25 based on the AIRS data
17 set as shown in Table 3-5 for the Detroit, MI; Chicago, IL; and Los Angeles, CA MS As.
18 However, data for analyzing the spatial variability of PM10_25 are more limited than for PM25;
19 therefore, fewer urban areas could be characterized in Appendix 3A (Figures 3A-28 to 3A-33).
20 Whereas PM2 5 concentrations were found to be highly correlated between sites in the Detroit, MI
21 MSA (Table 3-4), estimated PM10_25 concentrations are noticeably less well correlated. Likewise,
22 correlations of PM10_2 5 in the Chicago, IL MSA are also lower than those for PM2 5. However
23 correlations of PM10_2 5 concentrations between several pairs of sites in the Los Angeles-Long
24 Beach partial MSA are higher than those for PM2 5.
25 The interpretation of these results is not straightforward, as concentrations of PM10_25 are
26 generated by taking the difference between collocated PM2 5 and PM10 monitors. Consequently,
27 caution must be exercised when viewing them. Errors in the measurement of PM2 5 and PM10
28 may play a large role in reducing apparent correlations of PM10_25 such that collocated PM10_25
29 "measurements" may be expected to be poorly correlated (White, 1998). Indeed, several
30 estimates are negative. The possible causes of these errors are essentially the same as those
31 discussed in Section 3.2.1 with regard to the occurrence of PM2 5 to PM10 ratios greater than one.
April 2002 3-48 DRAFT-DO NOT QUOTE OR CITE
-------
Correlations of PM Exposure Indicators
Philadelphia, Summer, 1992-93, 7 Sites
1.0
0.9
0.8
•=• 0.7
.
o 0.6
1
O 0.5
o
;ro 0.4
g
O 0.3
0.2
0.1
0.0
A
I '\
I
PM
10
PM
10-2.5
Average r
PM
PM
2.5
10
0.90
0.86
A PM10_2.5 0.38
8 12 16 20 24 28 32
Distance Between Sites (km)
Not Significant, all other correlations significant (P<0.05)
Figure 3-19. Intersite correlation coefficients for PM2 5, PM10, and PM10_2 5.
Source: Wilson and Suh (1997).
1 There are also physical bases for expecting that PM10_2 5 concentrations may be more variable
2 than those for PM2 5. PM10.2 5 is mainly primary in origin, and its emissions are spatially and
3 temporally heterogenous. Similar considerations apply to primary PM25, but much of PM25 is
4 secondary, and sources of secondary PM are much less spatially and temporally variable. Dry
5 deposition rates of particles depend strongly on particle size. Whereas all particles may be
6 brought to the surface by turbulent motions in the atmosphere; gravitational settling becomes
7 more important with increasing particle size. Gravitational settling can effectively limit the
8 horizontal distance a particle can travel. For example, 10 //m Da particles suspended in a
April 2002
3-49
DRAFT-DO NOT QUOTE OR CITE
-------
1 hypothetical 1 km deep planetary boundary layer can be removed within a few hours, but 1 //m
2 Da particles can remain suspended in the atmosphere for up to 100 to 1,000 times longer before
3 being dry deposited. (Estimated atmospheric lifetimes were based on deposition velocities given
4 in Lin et al. [1994] for typical wind speeds.) The findings of larger correlations of PM10_25
5 between several site pairs in the Los Angeles basin and one other site pair in the St. Louis,
6 MO-IL MSA (cf. Figure 3A-17 and 3A-30) are anomalous in light of the discussion above.
7 However, these findings could have resulted from differences between the spatial and temporal
8 behavior of sources of PM25 and PM10_2 5 in these locations. Because of negative values, CODs
9 were not calculated.
10
11 PM Components
12 Three methods for comparing the chemical composition of aerosol databases obtained at
13 different locations and times were discussed by Wongphatarakul et al. (1998). Log-log plots of
14 chemical concentrations obtained at pairs of sampling sites accompanied by the coefficient of
15 divergence (COD) were examined as a way to provide an easily visualized means of comparing
16 two data sets2. Examples comparing downtown Los Angeles with Burbank and with
17 Riverside-Rubidoux are shown in Figures 3-20 and 3-21, respectively. As the composition of
18 two sampling sites become more similar, the COD approaches zero; as their compositions
19 diverge, the COD approaches one. Correlation coefficients calculated between components can
20 be used to show the degree of similarity between pairs of sampling sites.
21 In addition to calculating correlation coefficients for total mass or for individual
22 components, correlation coefficients for characterizing the spatial variation of the contributions
23 from given source types can also be calculated by averaging the correlation coefficients of the set
24 of chemical components that represent the source type. Correlation coefficients showing the
25 spatial relations among PM2 5 (total) and contributions from different source categories obtained
2The COD for two sampling sites is defined as follows:
-
P 7=1 tj •
where x^ represents the average concentration for a chemical component i at site j, j and k represent two sampling
sites, and p is the number of chemical components.
April 2002 3-50 DRAFT-DO NOT QUOTE OR CITE
-------
102
10°-
CO
£ 10-M
CD
io-2-
COD=0.099
T01
10
-3
so.
2-
nknown
-NO;
1 I I
T2 1C'1 10°
Downtown Los Angeles (|jg/m3)
102
Figure 3-20. PM2 5 chemical components in downtown Los Angeles and Burbank
(1986) have similar characteristics. The spread in the data is shown by
the bars.
Source: Wongphatarakul et al. (1998).
April 2002
3-51
DRAFT-DO NOT QUOTE OR CITE
-------
10
-3
10'2 10'1 10°
Downtown Los Angeles (|jg/m3)
Figure 3-21. Concentrations of PM25 chemical components in Rubidoux and downtown
Los Angeles (1986). The diagram shows a significant spread in the
concentrations for the two sites compared with downtown Los Angeles and
Burbank (Figure 3-20).
Source: Wongphatarakul et al. (1998).
April 2002
3-52
DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
5
9
10
11
12
13
at various sites in the South Coast Air Basin (SoCAB) Study are shown in Table 3-6.
In Wongphatarakul et al. (1998), crustal material (crustal), motor vehicle exhaust (mv), residual
oil emissions (residual oil), and secondary PM (sec) were considered as source categories.
Al, Si, Fe, and Ca were used as markers for crustal material (crustal). V and Ni were used as
markers for fuel oil combustion (residual oil). Pb, Br, and Mn were used as markers for motor
vehicle exhaust (mv), based on the lack of other, perhaps more suitable, tracers. NO3", NH4+, and
SO4"2 represent secondary PM components (sec). The average of the correlation coefficients of
marker elements within each source category are shown in Table 3-6. Values of rsec and rmv are
much higher than those for rcrustal and rresidual oil throughout the SoCAB, suggesting a more uniform
distribution of the contributions from secondary PM formation and automobiles than from crustal
material and localized stationary sources.
TABLE 3-6. CORRELATION COEFFICIENTS FOR SPATIAL VARIATION OF
PM2 5 MASS AND DIFFERENT SOURCES FOR PAIRS OF SAMPLING
SITES IN THE SOUTH COAST AIR BASIN (1986)
Hawthorne and Rubidoux
Long Beach and Rubidoux
Anaheim and Rubidoux
Downtown Los Angeles and Rubidoux
Burbank and Rubidoux
Hawthorne and Anaheim
Long Beach and Anaheim
Burbank and Anaheim
Downtown Los Angeles and Anaheim
Downtown Los Angeles and Hawthorne
Burbank and Hawthorne
Long Beach and Burbank
Long Beach and Hawthorne
Downtown Long Angeles and Long Beach
Downtown Los Angeles and Burbank
^total
-0.027
0.051
0.066
0.095
0.120
0.760
0.852
0.770
0.827
0.808
0.704
0.731
0.880
0.842
0.928
-'crustal
0.034
0.075
0.105
0.143
0.568
0.599
0.633
0.649
0.653
0.825
rsec
0.768
0.888
0.749
0.804
0.854
0.790
0.737
0.909
0.817
0.960
-'mv
0.492
0.504
0.579
0.556
0.669
0.688
0.714
0.861
0.719
0.871
-'residual oil
0.170
0.150
0.161
0.233
0.533
0.491
0.295
0.482
0.378
0.606
Source: Wongphatarakul etal. (1998).
April 2002
3-53
DRAFT-DO NOT QUOTE OR CITE
-------
1 Correlation coefficients in Philadelphia air for PM2 5 (total), crustal components (Al, Si, Ca,
2 and Fe), the major secondary component (sulfate), organic carbon (OC), and elemental carbon
3 (EC) are shown in Table 3-7, based on data obtained at four sites. Because these data were
4 obtained after Pb had been phased out of gasoline, a motor vehicle contribution could not be
5 estimated from the data. Pb also is emitted by discrete point sources, such as the Franklin
6 smelter. Concentrations of V and Ni were often beneath detection limits; so, the spatial
7 variability in PM due to residual oil combustion were not estimated. Sulfate in aerosol samples
8 collected in Philadelphia arises mainly from long-range transport from regionally dispersed
9 sources (Dzubay et al., 1988). This conclusion is strengthened by the high correlations in sulfate
10 between different monitoring sites and the uniformity in sulfate concentrations observed among
11 the sites. Widespread area sources (e.g., motor vehicle traffic) also may emit pollutants that are
12 correlated between sites provided that traffic patterns and emissions are similar throughout the
13 area under consideration.
14
15
TABLE 3-7. CORRELATION COEFFICIENTS FOR SPATIAL VARIATION OF
PM2 5 MASS AND DIFFERENT COMPONENTS FOR PAIRS OF SAMPLING
SITES IN PHILADELPHIA (1994)
Castor Ave. and Roxboro
Castor Ave. and NE Airport
Castor Ave. and Broad St.
Roxboro and NE Airport
Roxboro and Broad St.
NE Airport and Broad St.
rtot
0.92
0.93
0.93
0.98
0.95
0.95
-'crustal
0.52
0.47
0.57
0.67
0.90
0.69
rsec
0.98
0.99
0.99
0.98
0.98
0.99
roc
0.88
0.88
0.85
0.83
0.86
0.84
TEC
0.84
0.77
0.89
0.82
0.79
0.63
rPb
0.43
-0.07
0.11
0.20
0.47
0.11
Source: Pinto et al. (1995).
1 Landis et al. (2001) found relatively high correlations between PM2 5 (r = 0.97), sulfate
2 (r = 0.99), OC (r = 0.97), EC (r = 0.83), NaCl (r = 0.83), and nitrate (r = 0.83) measured at two
3 sites located several km apart in the Baltimore, MD area. Concentrations of crustal material
April 2002 3-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 (r = 0.63) and the sum of total metal oxides (r = 0.76) were not as well correlated. These results
2 are consistent with those for another eastern city, Philadelphia, PA, given in Table 3-7. The
3 results presented above for Philadelphia, PA; Baltimore, MD; and Los Angeles, CA, indicate that
4 secondary PM components are more highly correlated than primary components and may be
5 more highly correlated than total PM25. These results suggest that the correlation of PM
6 concentrations across an urban area may depend on the relative proportions of primary and
7 secondary components of PM at individual sites. Sampling artifacts affecting the measurement
8 of nitrate and organic carbon can obscure these relations and may depress correlations between
9 sites.
10 Kao and Friedlander (1995) examined the statistical properties of a number of PM
11 components in the South Coast Air Basin (Los Angeles area). They found that, regardless of
12 source type and location within their study area, the concentrations of nonreactive, primary
13 components of PM10 had approximately log-normal frequency distributions with constant values
14 of the geometric standard deviations (GSDs). However, aerosol constituents of secondary origin
15 (e.g., SO4"2, NH4+, and NO3") were found to have much higher GSDs. Surprisingly, the GSDs of
16 organic (1.87) and elemental (1.74) carbon were both found to be within 1 SD (0.14) of the mean
17 GSD (1.85) for nonreactive primary species, compared to GSD's of 2.1 for sulfate, 3.5 for
18 nitrate, and 2.6 for ammonium. These results suggest that most of the organic carbon seen in
19 ambient samples in the South Coast Air Basin was of primary origin. Pinto et al. (1995) found
20 similar results for data obtained during the summer of 1994 in Philadelphia. Further studies are
21 needed to determine if these relations are valid at other locations and to what extent the results
22 might be influenced by sampling artifacts such as the evaporation of volatile constituents during
23 or after sampling.
24 Very few studies have compared aerosol composition in urban areas to that in nearby rural
25 areas. One exception is Tanner and Parkhurst (2000), which indicates that sulfate constituted a
26 larger fraction of fine particle mass at rural sites in the Tennessee Valley PM25 monitoring
27 network than did organic carbon. For urban sites, the situation was largely reversed, with organic
28 carbon constituting a larger fraction of aerosol mass than sulfate. Systematic comparisons of
29 urban-rural differences in aerosol properties will be facilitated in the future with the
30 implementation of the national speciation network and the continued operation of the IMPROVE
31 network.
April 2002 3-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.3 SOURCES OF PRIMARY AND SECONDARY PARTICULATE
2 MATTER
3 Information about the nature and relative importance of sources of ambient PM is presented
4 in this section. Table 3-8 summarizes anthropogenic and natural sources for the major primary
5 and secondary aerosol constituents of fine and coarse particles. Major sources of each
6 constituent are shown in boldface type. Anthropogenic sources can be further divided into
7 stationary and mobile sources. Stationary sources include fuel combustion for electrical utilities,
8 residential space heating, and industrial processes; construction and demolition; metals, minerals,
9 and petrochemicals; wood products processing; mills and elevators used in agriculture; erosion
10 from tilled lands; waste disposal and recycling; and fugitive dust from paved and unpaved roads.
11 Mobile or transportation-related sources include direct emissions of primary PM and secondary
12 PM precursors from highway and off-highway vehicles and non-road sources. In addition to
13 fossil fuel combustion, biomass in the form of wood is burned for fuel. Vegetation is burned to
14 clear new land for agriculture and for building construction, to dispose of agricultural and
15 domestic waste, to control the growth of animal or plant pests, and to manage forest resources
16 (prescribed burning). Also shown are sources for precursor gases whose oxidation forms
17 secondary particulate matter. A description of the atmospheric chemical processes producing
18 secondary PM is given in Section 3.3.1.
19 In general, the sources of fine parti culate matter are very different from those for coarse
20 PM. Some of the mass in the fine size fraction has been formed during combustion from
21 material that has volatilized in combustion chambers and then recondensed before emission into
22 the atmosphere. By and large, however, most ambient PM2 5 has been formed in the atmosphere
23 from photochemical reactions involving precursor gases. PM formed by the first mechanism is
24 referred to as primary, and PM formed by the second mechanism is referred to as secondary.
25 PM10_2 5 is mainly primary in origin as it is produced by the abrasion of surfaces or by the
26 suspension of biological material. Because precursor gases undergo mixing during transport
27 from their sources, it is difficult to identify individual sources of secondary constituents of PM.
28 Transport and transformations of precursors can occur over distances of hundreds of kilometers.
29 The coarse PM constituents have shorter lifetimes in the atmosphere, so their effects tend to be
30 more localized. Only major sources for each constituent within each broad category shown at the
31 top of Table 3-8 are listed. Not all sources are equal in magnitude. Chemical characterizations
April 2002 3-56 DRAFT-DO NOT QUOTE OR CITE
-------
>
TABLE 3-8. CONSTITUENTS OF ATMOSPHERIC PARTICLES AND THEIR MAJOR SOURCES1
*— Sources
§ Primary (PM <2 . 5 ^m) Primary (PM >2 . 5 ^m) Secondary PM Precursors (PM <2 . 5 /j,m)
Aerosol
species Natural Anthropogenic Natural
SO4= Sea spray Fossil fuel combustion Sea spray
Sulfate
Anthropogenic Natural
— Oxidation of reduced sulfur
gases emitted by the oceans and
wetlands and SO2 and H2S
emitted by volcanism and forest
fires
Anthropogenic
Oxidation of SO2 emitted
from fossil fuel combustion
NO3-
Nitrate
Minerals
Oxidation of NO,, produced by
soils, forest fires, and lighting
Oxidation of NO,, emitted
from fossil fuel combustion
and in motor vehicle
exhaust
Erosion and Fugitive dust paved
re-entrainment and unpaved roads,
agriculture, and
forestry
Erosion and re-entrainment
Fugitive dust, paved
and unpaved road
dust, agriculture, and
forestry
~^/l
o
3>
'•Tj
H
6
o
2
0
H
O
c
o
~1
w
o
hrl
7s
H
W
NH4+
Ammonium
Organic
carbon (OC)
Elemental
carbon
(EC)
Metals
Bioaerosols
—
Wild fires
Wild fires
Volcanic
activity
Viruses and
bacteria
'Dash (-) indicates either very
—
Prescribed burning,
wood burning, motor
vehicle exhaust, and
cooking
Motor vehicle exhaust
wood burning, and
cooking
— —
— Tire and asphalt wear
and paved road dust
, — Tire and asphalt wear
and paved road dust
Emissions of NH3 from wild Emissions of NH3 from
animals, and undisturbed soil animal husbandry, sewage,
and fertilized land
Oxidation of hydrocarbons Oxidation of hydrocarbons
emitted by vegetation (terpenes, emitted by motor vehicles,
waxes) and wild fires prescribed burning, and
wood burning
— —
Fossil fuel combustion, Erosion, re-entrainment, — — —
smelting, and brake
wear
minor source or no known
and organic debris
Plant and insect fragments, —
pollen, fungal spores, and
bacterial agglomerates
source of component.
-------
1 of primary particulate emissions for a wide variety of natural and anthropogenic sources (as
2 shown in Table 3-8) were given in Chapter 5 of the 1996 PM AQCD. Summary tables of the
3 composition of source emissions presented in the 1996 PM AQCD and updates to that
4 information are provided in Appendix 3D. The profiles of source composition were based in
5 large measure on the results of various studies that collected signatures for use in source
6 apportionment studies.
7 Natural sources of primary PM include windblown dust from undisturbed land, sea spray,
8 and plant and insect debris. The oxidation of a fraction of terpenes emitted by vegetation and
9 reduced sulfur species from anaerobic environments leads to secondary PM formation.
10 Ammonium (NH4+) ions, which play a major role in regulating the pH of particles, are derived
11 from emissions of ammonia (NH3) gas. Source categories for NH3 have been divided into
12 emissions from undisturbed soils (natural) and emissions that are related to human activities
13 (e.g., fertilized lands, domestic and farm animal waste). There is ongoing debate about
14 characterizing emissions from wild fires (i.e., unwanted fire) as either natural or anthropogenic.
15 Wildfires have been listed in Table 3-8 as natural in origin, but land management practices and
16 other human actions affect the occurrence and scope of wildfires. For example, fire suppression
17 practices allow the buildup of fire fuels and increase the susceptibility of forests to more severe
18 and infrequent fires from whatever cause, including lightning strikes. Similarly, prescribed
19 burning is listed as anthropogenic, but can viewed as a substitute for wildfires that would
20 otherwise occur eventually on the same land.
21 The transformations that gaseous precursors to secondary PM formation undergo after
22 being emitted from the sources shown in Table 3-8 are described in Section 3.3.1. Aspects of the
23 transport of primary PM and secondary PM, including the transport of material from outside the
24 United States, are described in Section 3.3.2. A brief introduction to the deposition of particles is
25 also given in Section 3.3.2, and a more detailed discussion of deposition processes is presented in
26 Chapter 4. Methods to infer contributions from different source categories to ambient PM using
27 receptor models and the results of these modeling efforts are given in Section 3.3.3. Estimates of
28 emissions of primary PM and precursors to secondary PM from major sources are presented in
29 Section 3.3.4. A discussion of the uncertainties associated with these emissions is given in
30 Section 3.3.5.
31
April 2002 3-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.3.1 Chemistry of Secondary PM Formation
2 The precursors to secondary PM have natural and anthropogenic sources, just as primary
3 PM has natural and anthropogenic sources. Whereas the major atmospheric chemical
4 transformations leading to the formation of particulate nitrate and sulfate have been relatively
5 well understood; those involving the formation of secondary aerosol organic carbon are still
6 under investigation. A large number of organic precursors are involved; many of the kinetic
7 details still need to be determined; and many of the actual products of the oxidation of
8 hydrocarbons have yet to be identified.
9
10 Formation ofSulfates and Nitrates
11 A substantial fraction of the fine particle mass, especially during the warmer months of the
12 year, is secondary sulfate and nitrate formed as the result of atmospheric reactions. Such
13 reactions involve the gas phase conversion of SO2 to H2SO4 initiated by reaction with OH
14 radicals and aqueous-phase reactions of SO2 with H2O2, O3, or O2 (catalyzed by Fe and Mn).
15 These heterogeneous reactions may occur in cloud and fog droplets or in films on atmospheric
16 particles. NO2 can be converted to HNO3 by reaction with OH radicals during the day. At night,
17 NO2 also is oxidized to nitric acid by a sequence of reactions initiated by O3 that produce nitrate
18 radicals (NO3) and dinitrogenpentoxide (N2O5) as intermediates. Both H2SO4 and HNO3 react
19 with atmospheric ammonia (NH3). Gaseous NH3 reacts with gaseous HNO3 to form particulate
20 NH4NO3. Gaseous NH3 reacts with H2SO4 to form acidic HSO4 (in NH4 HSO4) as well as SO4'2
21 in (NH4)2SO4. In addition, acid gases such as SO2 and HNO3 may react with coarse particles to
22 form coarse secondary PM containing sulfate and nitrate. Examples include reactions with basic
23 compounds resulting in neutralization (e.g., CaCO3 + 2 HNO3 - Ca (NO3)2 + H2CO31) or with
24 salts of volatile acids resulting in release of the volatile acid (e.g., SO2 + 2NaCl + H2O - Na2SO3
25 +2HC1T).
26 If particulate NH4NO3 coagulates with an acidic sulfate particle (H2SO4 or HSO4), gaseous
27 HNO3 will be released, and the NH3 will increase the neutralization of the acidic sulfate. Thus,
28 in the eastern United States, where PM tends to be acidic, sulfate is usually a larger fraction of
29 PM mass than nitrate. However, in the western United States, where higher NH3 and lower SO2
30 emissions permit complete neutralization of H2SO4, the concentration of nitrate may be higher
31 than that of sulfate. As SO2 concentrations in the atmosphere in the eastern United States are
April 2002 3-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 reduced, the NH3 left in the atmosphere after neutralization of H2SO4 will be able to react with
2 HNO3 to form NH4NO3. Therefore, a reduction in SO2 emissions, especially without a reduction
3 in NOX emissions, could lead to an increase in NH4NO3 concentrations (West et al., 1999; Ansari
4 and Pandis, 1998). Thus, possible environmental effects of NH4NO3 are of interest for both the
5 western and eastern United States.
6 Chemical reactions of SO2 and NOX within plumes are an important source of H+, SO4"2,
7 and NO3. These conversions can occur by gas-phase and aqueous-phase mechanisms. In power-
8 plant or smelter plumes containing SO2 and NOX, the gas-phase chemistry depends on plume
9 dilution, sunlight, and volatile organic compounds either in the plume or in the ambient air
10 mixing into and diluting the plume. For the conversion of SO2 to H2SO4 in the gas-phase in such
11 plumes during summer midday conditions in the eastern United States, the rate typically varies
12 between 1 and 3% h"1 but in the cleaner western United States rarely exceeds 1% h"1. For the
13 conversion of NOX to FDSTO3, the gas-phase rates appear to be approximately three times faster
14 than the SO2 conversion rates. Winter rates for SO2 conversion are approximately an order of
15 magnitude lower than summer rates.
16 The contribution of aqueous-phase chemistry to particle formation in point-source plumes
17 is highly variable, depending on the availability of the aqueous phase (wetted aerosols, clouds,
18 fog, and light rain) and the photochemically generated gas-phase oxidizing agents, especially
19 H2O2 for SO2 chemistry. The in-cloud conversion rates of SO2 to SO4"2 can be several times
20 larger than the gas-phase rates given above. Overall, it appears that SO2 oxidation rates to SO4"2
21 by gas-phase and aqueous-phase mechanisms may be comparable in summer, but aqueous-phase
22 chemistry may dominate in winter. Further details concerning the chemistry of SO2 and NOX in
23 power plant plumes can be found in Hewitt (2001).
24 In the western United States, markedly higher SO2 conversion rates have been reported in
25 smelter plumes than in power plant plumes. The conversion occurs predominantly by a gas-
26 phase mechanism. This result is attributed to the lower NOX in smelter plumes. In power plant
27 plumes, NO2 depletes OH radicals and competes with SO2 for OH radicals.
28 In urban plumes, the upper limit for the gas-phase SO2 conversion rate appears to be about
29 5% h"1 under the more polluted conditions. For NO2, the rates appear to be approximately three
30 times faster than the SO2 conversion rates. Conversion rates of SO2 and NOX in background air
31 are comparable to the peak rates in diluted plumes. Neutralization of H2SO4 formed by SO2
April 2002 3-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 conversion increases with plume age and background NH3 concentration. If the NH3
2 concentrations are more than sufficient to neutralize H2SO4 to (NH4)2SO4, the HNO3 formed from
3 NOX conversions may be converted to NH4NO3.
4
5 Formation of Secondary Organic Particulate Matter (SOPM)
6 Atmospheric reactions involving volatile organic compounds such as alkanes, alkenes,
7 aromatics, cyclic olefins, and terpenes (or any reactive organic gas that contains at least seven
8 carbon atoms) yield organic compounds with low saturation vapor pressures at ambient
9 temperature. Such reactions may occur in the gas phase, in fog or cloud droplets (Graedel and
10 Goldberg, 1983; Faust, 1994), or possibly in aqueous aerosols (Aumont et al., 2000). Reaction
11 products from the oxidation of reactive organic gases also may nucleate to form new particles or
12 condense on existing particles to form secondary organic PM (SOPM). Organic compounds with
13 two double bounds may react to form dicarboxylic acids, which, with four or more carbon atoms,
14 also may condense. Both biogenic and anthropogenic sources contribute to primary and
15 secondary organic particulate matter (Grosjean, 1992; Hildemann et al., 1996; Mazurek et al.,
16 1997; Schauer et al., 1996). Oxalic acid was the most abundant organic acid found in PM2 5 in
17 California (Poore, 2000).
18 Although the mechanisms and pathways for forming inorganic secondary particulate matter
19 are fairly well known, those for forming SOPM are not as well understood. Ozone and the OH
20 radical are thought to be the major initiating reactants. However, HO2 and NO3 radicals also may
21 initiate reactions and organic radicals may be nitrated by HNO2, HNO3, or NO2. Pun et al. (2000)
22 discuss formation mechanisms for highly oxidized, multifunctional organic compounds. The
23 production of such species has been included in a photochemical model by Aumont et al. (2000),
24 for example. Understanding the mechanisms of formation of secondary organic PM is important
25 because SOPM can contribute in a significant way to ambient PM levels, especially during
26 photochemical smog episodes. Experimental studies of the production of secondary organic PM
27 in ambient air have focused on the Los Angeles Basin. Turpin and Huntzicker (1991, 1995) and
28 Turpin et al. (1991) provided strong evidence that secondary PM formation occurs during periods
29 of photochemical ozone formation in Los Angeles and that as much as 70% of the organic carbon
30 in ambient PM was secondary in origin during a smog episode in 1987. Schauer et al. (1996)
April 2002 3-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 estimated that 20 to 30% of the total organic carbon PM in the <2.1 //m size range in the
2 Los Angeles airshed is secondary in origin on an annually averaged basis.
3 Pandis et al. (1992) identified three mechanisms for formation of SOPM: (1) condensation
4 of oxidized end-products of photochemical reactions (e.g., ketones, aldehydes, organic acids, and
5 hydroperoxides), (2) adsorption of semivolatile organic compounds (SVOC) onto existing solid
6 particles (e.g., polycyclic aromatic hydrocarbons), and (3) dissolution of soluble gases that can
7 undergo reactions in particles (e.g., aldehydes). The first and third mechanisms are expected to
8 be of major importance during the summertime when photochemistry is at its peak. The second
9 pathway can be driven by diurnal and seasonal temperature and humidity variations at any time
10 of the year. With regard to the first mechanism, Odum et al. (1996) suggested that the products
11 of the photochemical oxidation of reactive organic gases are semivolatile and can partition
12 themselves onto existing organic carbon at concentrations below their saturation concentrations.
13 Thus, the yield of SOPM depends not only on the identity of the precursor organic gas but also
14 on the ambient levels of organic carbon capable of absorbing the oxidation products.
15 Haagen-Smit (1952) first demonstrated that hydrocarbons irradiated in the presence of NOX
16 produce light scattering aerosols. The aerosol forming potentials of a wide variety of individual
17 anthropogenic and biogenic hydrocarbons were compiled by Pandis et al. (1992), based mainly
18 on estimates made by Grosjean and Seinfeld (1989) and data from Pandis et al. (1991) for
19 p-pinene and from Izumi and Fukuyama (1990) for aromatic hydrocarbons. Zhang et al. (1992)
20 examined the oxidation of a-pinene. Pandis et al. (1991) found no aerosol products formed in
21 the photochemical oxidation of isoprene, although they and Zhang et al. (1992) found that the
22 addition of isoprene to reaction mixtures increased the reactivity of the systems studied. Further
23 details about the oxidation mechanisms and secondary organic PM yields from various reactive
24 organic gases are given in the above studies. Estimates of the production rate of secondary
25 organic PM in the Los Angeles airshed are provided in the 1996 PM AQCD (U.S. Environmental
26 Protection Agency, 1996).
27 More recently, Odum et al. (1997a,b) have found that the aerosol formation potential of
28 whole gasoline vapor can be accounted for solely by summing the contributions of the individual
29 aromatic compounds in the fuel. In general, data for yields for secondary organic PM formation
30 can be broken into two distinct categories. The oxidation of toluene and aromatic compounds
31 containing ethyl or propyl groups (i.e., ethylbenzene, ethyltoluene, n-propylbenzene) produced
April 2002 3-62 DRAFT-DO NOT QUOTE OR CITE
-------
1 higher yields of secondary organic PM than did the oxidation of aromatic compounds containing
2 two or more methyl groups (i.e., xylenes, di-, tri-, tetra-methylbenzenes). Yields in the first
3 group ranged from about 7 to 10%; and in the second group, they ranged from 3 to 4% for
4 organic carbon concentrations between 13 and 100 //g/m3. Reasons for the differences in
5 secondary organic PM yields found between the two classes of compounds are not clear.
6 There have been a few recent studies that have examined the composition of secondary
7 organic PM. Edney et al., (2001) carried out a smog chamber study to investigate the formation
8 of multi-functional oxygenates from photooxidation of toluene. The experiments were carried
9 out by irradiating toluene/propylene/NOx/air mixtures in a smog chamber operated in the
10 dynamic mode and analyzing the collected aerosol by positive chemical ionization GC-MS after
11 derivatization of the carbonyl oxidation products. The results of the GC-MS analyses were
12 consistent with the formation of semivolatile multi-functional oxygenates, including hydroxy
13 diones as well as tri ones, tetraones, and pentaones. The authors also suggested that some of these
14 compounds could be present in SOPM in the form of polymers.
15 Jang and Kamens (200la) employed a number of analytical approaches, including GC-MS
16 detection of volatile derivatives of carbonyl, hydroxy, and acid compounds in SOPM formed in
17 the irradiation of toluene/NOx mixtures. A wide range of substituted aromatics, nonaromatic
18 ring-retaining and ring-opening products were detected. Newly identified ring opening
19 oxycarboxylic acids detected included: glyoxylic acid; methylglyoxylic acid; 4-oxo-2-butenoic
20 acid; oxo-C5-alkenoic acids; dioxopentenoic acids; oxo-C7-alkadienoic acids; dioxo-C6-alkenoic
21 acids; hydroxydioxo-C7-alkenoic acids; and hydroxytrioxo-C6-alkanoic acids. Other newly
22 identified compounds included methylcyclohexenetriones; hydroxymethylcyclohexenetriones;
23 2-hydroxy-3-penten-l,5-dial, hydroxyoxo-C6-alkenals; hydroxy-C5-triones, hydroxydioxo-C7-
24 alkenals; and hydroxy-C6-tetranones. Included among these compounds were a number of the
25 hydroxy polyketones detected by Edney et al., (2001). Recent laboratory and field studies
26 support the concept that nonvolatile and semivolatile oxidation products from the photooxidation
27 of biogenic hydrocarbons contribute significantly to ambient PM concentrations in both urban
28 and rural environments. The oxidation of a variety of biogenic hydrocarbons emitted by trees
29 and plants, such as terpenes (or-pinene, /6-pinene, 4J-carene, sabinene, or-terpinene, y-terpinene,
30 terpinolene, myrcene, and ocimene) and sesequiterpenes (/S-caryophyllene and or-humulene)
31 could form SOPM. Vegetation also emits oxygenated organic compounds such as alcohols,
April 2002 3-63 DRAFT-DO NOT QUOTE OR CITE
-------
1 acetates, aldehyde, ketones, ethers, and esters (Winer et al., 1992). However, their contribution
2 to SOPM, remains uncertain. Hoffmann et al. (1997) found SOPM yields of -5% for open-chain
3 biogenic hydrocarbons such as ocimene and linalool; 5 to 25% for monounsaturated cyclic
4 monoterpenes such as a-pinene; A-3 carene and terpinene-4-ol; and -40% for a cyclic
5 monoterpene with an endocyclic and an exocyclic double bond such as d-limonene. Secondary
6 organic PM yields of close to 100% were observed during the photochemical oxidation of one
7 sesquiterpene, trans-caryophyllene. These results were all obtained for initial hydrocarbon
8 mixing ratios of 100 ppb, which are much higher than found in the atmosphere.
9 Kamens et al. (1999) observed SOPM yields of 20 to 40% for a-pinene. Using information
10 on the composition of secondary PM formed from a-pinene (Jang and Kamens, 1999), they were
11 able to calculate formation rates with a kinetic model including formation mechanisms for
12 O3 + a-pinene reaction products.
13 Griffin et al. (1999) introduced the concept of incremental aerosol reactivity, the change in
14 the secondary organic aerosol mass produced (in //g/m3) per unit change of parent organic reacted
15 (in ppb), as a measure of the aerosol-forming capability of a given parent organic compound in a
16 prescribed mixture of other organic compounds. They measured the incremental aerosol
17 reactivity for a number of aromatic and biogenic compounds for four initial mixtures.
18 Incremental aerosol reactivity ranged from 0.133 to 10.352 //gm"3 ppb"1 and varied by almost a
19 factor of two depending on the initial mixture.
20 A number of multifunctional oxidation products produced by the oxidation of biogenic
21 hydrocarbons have been identified in laboratory studies (Yu et al., 1998; Glasius et al., 2000;
22 Chri staffer sen et al., 1998; Koch et al., 2000; and Leach et al., 1999). Many of these compounds
23 have subsequently been identified in field investigations (Yu et al., 1999; Kavouras et al., 1998,
24 1999a,b; Pio et al., 2001; and Castro et al., 1999). Most studies of the formation of secondary
25 organic aerosol formation from terpenes have focused on their reactions with ozone. There have
26 been many fewer studies dealing with the oxidation of terpenes initiated by OH radicals. Larson
27 et al. (2001) found that the major aerosol products produced ultimately from the reaction of OH
28 radicals with mono-terpenes with endocyclic double bonds (a-pinene, 3-carene) were C10
29 kato-carboxylic acids (such as pinonic and caronic acids); whereas the major products from the
30 oxidation of mono-terpenes with exocyclic double bonds (p-pinene) were C9-dicarboxylic acids
31 (such as pinic acid), and the major product from the oxidation of limonene (which has both
April 2002 3-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 endo- and exocyclic double bonds) was 3-acetyl-6-oxo-heptanal (Kato-limonaldehyde). A large
2 number of related aldehydes, ketones and acids were also found in their experiments. However,
3 the total yields of condensable products are much lower than for the corresponding reactions with
4 ozone. For example, yields of C9-dicarboxylic acids, C10-hydroxy-keto-carboxylic acids, and
5 C10-hydroxy-Kato-aldehydes from the reaction of ozone with mono-terpenes with endocyclic
6 double bonds ranged form 3% to 9%; whereas they ranged only from 0.4 to 0.6% in the reaction
7 with OH radicals. Likewise, the reaction of mono-terpenes with exocyclic double bonds with
8 ozone produced much higher yields (1% to 4%) of C8- and C9-dicarboxylic acids than did their
9 reaction with OH radicals (0.2% to 0.3%). Apart from the complex products noted above, it
10 should be remembered that much simpler products, such as formaldehyde and formic acid, are
11 also formed in much larger yields form the same reactants (e.g., Winterhalter et al., 2000).
12 Compounds such as these also contribute to the formation of secondary organic aerosol
13 according to the mechanisms given in Pandis et al. (1992) and mentioned earlier in this section.
14 It is worth noting that the dicarboxylic acids and hydroxy-Kato-carboxylic acids have very
15 low vapor pressures and may act as nucleating species in OH- and O3- terpene reactions (Larson
16 et al., 2001). The rate coefficient for reaction of a-pinene with OH radicals is approximately a
17 factor of 106 greater than for its reaction with O3, based on data given in Atkinson (1994). The
18 daytime average concentration of O3 is typically a factor of 106 greater than that for OH radicals
19 in polluted boundary layers; whereas the above mentioned yields of aerosol products are roughly
20 a factor often greater in the O3-initiated reaction than in the corresponding OH radical reaction.
21 The foregoing suggests that the O3-initiated reaction may be more important than the OH
22 initiated reaction for the formation of aerosol products. Because ambient ozone is present at
23 night and it penetrates indoors, new particles may also be generated under these conditions.
24 For example, Wainman et al. (2000) found that ozone can react with limonene released by air
25 fresheners in indoor environments to produce substantial quantities of submicron particles. The
26 corresponding reaction involving OH radicals at night and in indoor environments is expected to
27 be negligible by comparison because of the very low OH concentrations present in these
28 environments. Although much progress has been made in determining the importance of
29 anthropogenic and biogenic hydrocarbons for the formation of secondary organic PM, further
30 investigations are needed to accurately assess their overall contributions to PM25 concentrations.
April 2002 3-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 Reactions of organic compounds either in particles or on the surface of particles have only
2 recently come under study. Tobias and Ziemann (2000) reported evidence for the formation of
3 relatively stable low volatility peroxy hemiacetals from reactions of hydroperoxides with
4 aldehydes and ketones on the surface of secondary organic particles. Shortly after the publication
5 of these results, Jang and Kamens (200 la) suggested, based on results of their laboratory
6 investigations of SOPM formation from irradiation of toluene/propylene/NOx /air mixtures, that
7 carbonyls and hydroxy compounds (either within or on the surface of aromatic SOPM) could
8 react together to form larger and less volatile hemiacetals and acetals. They also proposed that
9 dissolved carbonyls could undergo further reactions leading to the formation of a polymer, a
10 mechanism that has also been suggested by Edney et al. (2001). Each of these mechanisms that
11 also could be catalyzed by the presence of acids involves converting, through heterogenous
12 reactions, volatile compounds into much less volatile compounds, a mechanism that could
13 contribute to SOPM yields in aromatic and possibly biogenic systems.
14 As a first step in addressing these issues, Jang and Kamens (2001a) carried out a series of
15 laboratory screening experiments to assess whether volatile carbonyl compounds absorbed into
16 particles could undergo further chemical reactions forming low vapor pressure compounds.
17 Experiments were carried out whereby carbonyls were introduced in Teflon bags in the dark in
18 the presence of a seed aerosol containing either ammonium sulfate or a mixture of ammonium
19 sulfate and sulfuric acid. The increase in the aerosol volume was then measured using a scanning
20 mobility particle sizer. The carbonyls employed for the study included glyoxal, hexanal, and
21 octanal. Increased organic aerosol yields were found in the presence of the ammonium sulfate
22 seed aerosol for each of the carbonyls, with the highest yield being found for octanol followed in
23 decreasing order by glyoxal and then octanal. The presence of the acidified sulfate salt
24 significantly increased the yields even further. In a number of other experiments, 1-decanol was
25 added to the carbonyl-aerosol system to investigate the possible formation of hemiacetals and/or
26 acetals. Again, the volume of aerosol increased in both the presence of ammonium sulfate
27 aerosol and the acidified salt with a significantly larger yield found in the presence of acidity.
28 To explain their findings for acid-catalyzed carbonyl reactions, Jang and Kamens (2001a,b)
29 proposed a chemical mechanism in which the dissolved carbonyl first undergoes a protonization
30 reaction forming an adduct that can react with water to form its hydrate (1,1-dihydroxy gem-
31 diol). The adducts can then react with OH groups of the gem-diol forming higher molecular
April 2002 3-66 DRAFT-DO NOT QUOTE OR CITE
-------
1 weight and less volatile dimers that are subject to further reactions. In principal, this process,
2 which the authors refer to as a "zipping reaction" can lead to the formation of polymers.
3 However, because the individual reactions are reversible, the process can also be reversed by an
4 unzipping reaction. The zipping process could serve as an important mechanism for SOPM
5 formation by converting volatile oxidation products including glyoxal and methyl glyoxal into
6 low volatility compounds. On the other hand, the unzipping process that could take place during
7 the workup of the aerosol samples could be responsible for the detection of high volatile
8 oxidation products in SOPM, including glyoxal and methyl glyoxal that has been reported by
9 Edney et al. (2001), Cocker et al. (2001), and Jang and Kamens (2001a). While these processes
10 may take place in the absence of significant acidity, the experimental results suggest the
11 processes are likely enhanced by acid-catalyzed reactions.
12 Sampling and characterizing PM in the ambient atmosphere and in important
13 microenvironments is required to address important issues in exposure, toxicology, and
14 compliance. Currently, it is not possible to fully quantify the concentration, composition, or
15 sources of the organic components. Many of the secondary organic aerosol components are
16 highly oxidized, difficult to measure, multifunctional compounds. Additional laboratory studies
17 are needed to identify such compounds, strategies need to be developed to sample and measure
18 such compounds in the atmosphere, and models of secondary organic aerosol formation need to
19 be improved and added to air quality models in order to address issues related to human
20 exposure.
21 A high degree of uncertainty is associated with all aspects of the calculation of secondary
22 organic PM concentrations. This is compounded by the volatilization of organic carbon from
23 filter substrates during and after sampling as well as potential positive artifact formation from the
24 absorption of gaseous hydrocarbon on quartz filters. Significant uncertainties always arise in the
25 interpretation of smog chamber data because of wall reactions. Limitations also exist in
26 extrapolating the results of smog chamber studies to ambient conditions found in urban airsheds
27 and forest canopies. Concentrations of terpenes and NOX are much lower in forest canopies
28 (Altshuller, 1983) than the levels commonly used in smog chamber studies. The identification of
29 aerosol products of terpene oxidation has seldom been a specific aim of field studies, making it
30 difficult to judge the results of model calculations of secondary organic PM formation.
31 Uncertainties also arise because of the methods used to measure biogenic hydrocarbon emissions.
April 2002 3-67 DRAFT-DO NOT QUOTE OR CITE
-------
1 Khalil and Rasmussen (1992) found much lower ratios of terpenes to other hydrocarbons (e.g.,
2 isoprene) in forest air than were expected based on their relative emissions strengths and rate
3 coefficients for reaction with OH radicals and O3. They offered two explanations: (1) either the
4 terpenes were being removed rapidly by some heterogeneous process or (2) emissions were
5 enhanced artificially by feedbacks caused by the bag enclosures they used. If the former
6 consideration is correct, then the production of aerosol carbon from terpene emissions could be
7 substantial; if the latter is correct, then terpene emissions could have been overestimated by the
8 techniques used.
9
10 3.3.2 The Long-Range Transport of Particulate Matter from Outside the
11 United States
12 Apart from sources within the continental United States, particulate matter can be brought
13 in by long-range transport from sources outside the United States. For example, the transport of
14 PM from uncontrolled biomass burning in Central America and southern Mexico resulted in
15 anomalously high PM levels observed in southern Texas and generally elevated PM
16 concentrations throughout the entire central and southeastern United States during the spring and
17 early summer of 1998. Windblown dust from individual dust storms in the Sahara desert has
18 been observed in satellite images as plumes crossing the Atlantic Ocean and reaching the
19 southeast coast of the United States (e.g., Ott et al., 1991). Dust transport from the deserts of
20 Asia across the Pacific Ocean also occurs (Prospero, 1996). Most dust storms in the deserts of
21 China occur in the spring following the passage of strong cold fronts after the snow has melted
22 and before a surface vegetation cover has been established. Strong winds and unstable
23 conditions result in the rapid transport of dust to altitudes of several kilometers, where it is
24 transported by strong westerly winds out over the Pacific Ocean (Duce, 1995). Satellite images
25 were used to track the progress of a dust cloud from the Gobi desert to the northwestern United
26 States during the spring of 1998 (Husar et al., 2000).
27 Satellite images obtained at visible wavelengths cannot track mineral dust across the
28 continents because of a lack of contrast between the plume and the underlying surface. Other
29 means must be used to track the spread of North African dust through the eastern United States.
30 Perry et al. (1997) used two criteria (PM25 soil concentration > 3 //g/m3 and Al/Ca > 3.8) to
31 distinguish between soil of local origin from soil originating in North Africa in characterizing the
April 2002 3-68 DRAFT-DO NOT QUOTE OR CITE
-------
1 sources of PM in aerosol samples collected in the IMPROVE (Interagency Monitoring of
2 Protected Visual Environments) network. North African dust has been tracked as far north as
3 Illinois (Gatz and Prospero, 1996) and Maine (Perry et al., 1997). The analysis of Perry et al.
4 (1997) indicates that incursions of Saharan dust into the continental United States have occurred,
5 on average, about three times per year from 1992 to 1995. These events persist for about 10 days
6 on average, principally during the summer. Large scale dust events typically cover from 15 to
7 30% of the area of the continental United States and result in increases of PM2 5 levels of
8 8.7 ± 2.3 //g/m3 throughout the affected areas, with mean maximum dust contributions of
9 19.7 ± 8.4 //g/m3 during these events and a peak contribution of 32 //g/m3 to 24-h average PM 2 5
10 levels.
11 As can be expected, the frequency of dust events is highest in the southeastern United
12 States. About half of these events are observed only within the state of Florida, and these events
13 are associated with dense hazes in Miami (Figure 3-22) during the summer (Prospero et al.,
14 1987). North African dust is the dominant aerosol constituent in southern Florida during the
15 summer; whereas soil dust constitutes only a minor fraction of PM during the remainder of the
16 year (Prospero, 1999). Approximately one-third to one-half of the mass of the particles reaching
17 southern Florida have aerodynamic diameters less than 2.5 micrometers (Prospero et al., 2001).
18 During episodes when daily total dust concentrations ranged up to 100 //g/m3, it can be seen that
19 daily PM25 concentrations of up to 50 //g/m3 could have resulted in Miami, FL.
20 Husar et al. (2000) documented the transport of dust from the Gobi and Taklimakan deserts
21 to North America during April 1998. The PM10 concentration averaged over 150 stations in
22 Washington, Oregon, California, Nevada, and Idaho reporting data to AIRS was 65 //g/m3
23 between April 26 and May 1, compared to about 20 //g/m3 during the rest of April and May.
24 Data from several networks indicated that PM10 concentrations were over 100 //g/m3 in central
25 British Columbia, Washington State, and Oregon. The highest PM concentrations observed were
26 120 //g/m3 for PM10 and 50 //g/m3 for PM2 5 at Chilliwack Airport in northwestern Washington
27 State (Figure 3-23). Aircraft measurements made over the northwestern United States were
28 consistent with a mass median diameter of the dust being between 2 and 3 //m.
29 Desert dust deposited over oceans provides nutrients to marine ecosystems (Savoie and
30 Prospero, 1980). Desert dust deposited on nutrient depleted soils also provides nutrients, as in
31 Hawaiian rain forests (Chadwick et al., 1999). Microorganisms, including various species and
April 2002 3-69 DRAFT-DO NOT QUOTE OR CITE
-------
Year
Figure 3-22. Monthly average Saharan dust components of the aerosol sampled in Miami,
FL, from 1974 to 1996. Approximately one-third to one-half of Saharan dust
is in the PM2 5 size range.
Source: Prospero (1999).
1 genera of fungi and bacteria, have been found attached to African dust particles in the U.S.
2 Virgin Islands (Griffin et al., 2001). The fungus, Aspergillus sydowii, which has been connected
3 to the death of coral reefs, has been identified in air samples collected in the Caribbean during
4 African dust transport events (Smith et al., 1996; Shinn et al., 2000). Measurements of the
5 composition of Saharan dust in Miami indicate enhancements of nitrate, non-sea-salt sulfate,
6 ammonium, and trace metals over concentrations expected for clean marine air, suggesting
7 pollution emitted in Europe and North Africa as sources (Prospero, 1999). It is likely that many
8 other constituents will be found associated with dust from outside North America as more
9 measurements are made. It should be noted that, as North African dust and associated material
10
April 2002
3-70
DRAFT-DO NOT QUOTE OR CITE
-------
140
120-
^100
O)
o 80
§ 60
o
O
40-
20
0
4/25/98 4/27/98 4/29/98 5/1/98 5/3/98 5/5/98
Figure 3-23. PM2 5 and PM10 concentrations measured at Chilliwack Airport, located
in northwestern Washington State, just before and during the Asian
desert dust episode of April and May 1998.
Source: U.S. EPA Aerometric Information Retrieval System (AIRS).
1 are transported northward through the United States during the summer, they are added to the
2 mixture of primary and secondary PM generated domestically.
3 Biomass burning for agricultural purposes occurs normally during the spring of each year in
4 Central America and southern Mexico. During the spring of 1998, fires burned uncontrollably
5 because of abnormally hot and dry conditions associated with the intense El Nino of 1997 to
6 1998. PM10 concentrations observed in the southern Rio Grande Valley were elevated
7 substantially during the passage northward of the biomass burning plume produced by these fires
8 as shown in Figure 3-24. Elevated PM10 concentrations also were found as far north as St. Louis,
9 MO (Figure 3-25). As can be seen from Figure 3-24 and Figure 3-25, the elevations in PM
April 2002
3-71
DRAFT-DO NOT QUOTE OR CITE
-------
600
5/5 5/6 5/7 5/8 5/9 5/10 5/11 5/12 5/13 5/14 5/15 5/16 5/17 5/18 5/19 5/20 5/21 5/22 5/23 5/24 5/25 5/26 5/27 5/28
Time (days)
Figure 3-24. Time series of 24-h average PM10 concentrations observed in the Rio
Grande Valley during May 1998.
Source: U.S. EPA Aerometric Information Retrieval System (AIRS).
IE
180-
160-
140-
120-
100-
80-
60-
40-
20
0
Smoke
Event
PM10 24 hr Standard
PM25 24 hr
5 6 7 8 9 10 11 121314151617
May 1998
Figure 3-25. PM10 concentrations observed in St. Louis, MO, during May 1998.
Source: U.S. EPA Aerometric Information Retrieval System (AIRS).
April 2002
3-72
DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations were limited in duration. Uncontrolled wildfires occur in the United States every
2 year, but their effects on air quality throughout the United States still need to be evaluated
3 systematically. These fires can be widespread, and the frequency of their occurrence can vary
4 markedly from year to year. For example, approximately 26,000 km2 were consumed during
5 2000, but only a small fraction of this area was burnt during 2001 in the western United States.
6 Wildfires also occur in the boreal forests of northwestern Canada. Wotawa and Trainer (2000)
7 suggested that the plume from fires occurring in the Northwest Territories of Canada in early
8 July 1995 may have extended throughout most of the eastern United States, resulting in elevated
9 levels of CO and ozone. Simple scaling of their calculated excess CO concentrations because of
10 the fires, by the ratio of emission factors of PM2 5 to CO, indicates that the excess PM2 5
11 concentrations in the plume may have ranged from about 5 //g/m3 in the Southeast and increasing
12 to close to 100 //g/m3 in the northern Plains States.
13
14 3.3.3 Source Contributions to Ambient PM Determined by Receptor Models
15 Receptor models are perhaps the primary means used to estimate the contributions of
16 different source categories to PM concentrations at individual monitoring sites. Dispersion
17 models (i.e., three-dimensional chemistry and transport models) are formulated in a prognostic
18 manner (i.e., they attempt to predict species concentrations using a tendency equation that
19 includes terms based on emissions inventories, atmospheric transport, chemical transformations,
20 and deposition). Receptor models are diagnostic in their approach (i.e., they attempt to derive
21 source contributions based either on ambient data alone or in combination with data from the
22 chemical composition of sources). These methods have the advantage that they do not invoke all
23 of the uncertainties inherent in emissions inventories or in parameterizing atmospheric transport
24 processes in grid point models.
25 There are two main approaches to receptor modeling. Receptor models such as the
26 chemical mass balance (CMB) model (Watson et al., 1990a) relate source category contributions
27 to ambient concentrations based on analyses of the composition of ambient particulate matter and
28 source emissions samples. This technique has been developed for apportioning source categories
29 of primary parti culate matter and was not formulated to include the processes of secondary
30 parti culate matter formation. In the second approach, various forms of factor analysis are used,
31 which rely on the analysis of time series of compositional data from ambient samples to derive
April 2002 3-73 DRAFT-DO NOT QUOTE OR CITE
-------
1 both the composition of sources and the source contributions. Standard approaches such as
2 factor analysis or Principal Component Analysis (PCA) can apportion only the variance and not
3 the mass in an aerosol composition data set. The other techniques described below, PMF and
4 UNMIX do apportion mass, however. Positive matrix factorization (PMF) is a recently
5 developed multivariate technique (Paatero and Tapper, 1993 and 1994) that overcomes many of
6 the limitations of standard techniques, such as principal components analysis (PCA), by allowing
7 for the treatment of missing data and data near or below detection limits. This is accomplished
8 by weighting elements inversely according to their uncertainties. Standard methods such as PCA
9 weight elements equally regardless of their uncertainly. Solutions also are constrained to yield
10 non-negative factors. Both the CMB and the PMF approaches find a solution based on least
11 squares fitting and minimize an object function. Both methods provide error estimates for the
12 solutions based on estimates of the errors in the input parameters. It should be remembered that
13 the error estimates often contain subjective judgments. For a complete apportionment of mass,
14 all of the major sources affecting a monitoring site must be sampled for analysis by CMB;
15 whereas there is no such restriction in the use of PMF.
16 Among other approaches, the UNMIX model takes a geometric approach that exploits the
17 covariance of the ambient data to determine the number of sources, the composition and
18 contributions of the sources, and the uncertainties (Henry, 1997). A simple example may help
19 illustrate the approach taken by UNMIX. For example, in a two-element scatter plot of ambient
20 Al and Si, a straight line and a high correlation for Al versus Si can indicate a single source for
21 both species (soil), while the slope of the line gives information on the composition of the soil
22 source. In the same data set, iron may not plot on a straight line against Si, indicating other
23 sources of Fe in addition to soil. More importantly, the Fe-Si scatter plot may reveal a lower
24 edge. The points defining this edge represent ambient samples collected on days when the only
25 significant source of Fe was soil. Success of the UNMIX model hinges on the ability to find
26 these "edges" in the ambient data from which the number of sources and the source compositions
27 are extracted. UNMIX uses principal component analysis to find edges in m-dimensional space,
28 where m is the number of ambient species. The problem of finding edges is more properly
29 described as finding hyperplanes that define a simplex. The vertices at which the hyperplanes
30 intersect represent pure sources from which source compositions can be determined. However,
31 there are measurement errors in the ambient data that "fuzz" the edges making them difficult to
April 2002 3-74 DRAFT-DO NOT QUOTE OR CITE
-------
1 find. UNMIX employs an "edge-finding" algorithm to find the best edges in the presence of
2 error. UNMIX does not make explicit use of errors or uncertainties in the ambient
3 concentrations, unlike the methods outlined above. This is not to imply that the UNMIX
4 approach regards data uncertainty as unimportant, but rather that the UNMIX model results
5 implicitly incorporate error in the ambient data. The underlying philosophy is that the
6 uncertainties are often unquantifiable, and hence it is best to make no a priori assumptions about
7 what they are.
8 In addition to chemical speciation data, Norris et al. (1999) showed that meteorological
9 indices could prove useful in identifying sources of particulate matter that are responsible for
10 observed health effects (specifically asthma) associated with exposure to parti culate matter.
11 They examined meteorology associated with elevated pollution events in Spokane and Seattle,
12 WA, and identified a "stagnation index" that was associated with low wind speeds and increases
13 in concentrations of combustion-related pollutants. Their factor analysis also identified a
14 meteorological index (low relative humidity and high temperatures) that was associated with
15 increases in soil-derived paniculate matter, as well as a third factor (low temperatures and high
16 relative humidity) that was associated with increases in concentrations of parti culate sulfate and
17 nitrate species (Norris, 1998).
18 Ondov (1996) examined the feasibility of using sensitive isotopic and elemental tracer
19 materials to determine the contributions of petroleum-fueled sources of PM10 in the San Joaquin
20 Valley, CA. Costs of these experiments are affected not only by the tracer materials cost, but
21 also by the sensitivities of the analytical methods for each, as well as the background levels of the
22 tracers. Suarez et al. (1996) used iridium as a tracer to tag emissions from diesel-burning
23 sanitation trucks in Baltimore and determined the size distribution of soot from the trucks.
24 A number of specialty conference proceedings, review articles, and books have been
25 published that provide greater detail about source category apportionment receptor models then
26 described in the 1996 PM AQCD. A review of the various methods used to apportion PM in
27 ambient samples among its source categories was given in Section 5.5.2 of the 1996 PM AQCD.
28 The collection of the source category characterization profiles shown in Appendix 3D has been
29 motivated in many cases by the need to use them in receptor modeling applications.
30 The results of several source apportionment studies are discussed in this section to provide
31 an indication of the relative importance of different sources of parti culate matter across the
April 2002 3-75 DRAFT-DO NOT QUOTE OR CITE
-------
1 United States. First, results obtained mainly by using the chemical mass balance (CMB)
2 approach for estimating contributions to PM2 5 from different source categories at monitoring
3 sites in the United States are discussed and presented in Table 3-9. More recent results using the
4 PMF approach are included for Phoenix, AZ. Results obtained at a number of monitoring sites in
5 the central and western United States by using the CMB model for PM10 are shown in
6 Table 3-10. The sampling sites represent a variety of different source characteristics within
7 different regions of Arizona, California, Colorado, Idaho, Illinois, Nevada, and Ohio. Definitions
8 of source categories also vary from study to study. The results of the PM10 source apportionment
9 studies were given in the 1996 PM AQCD and are presented here to allow easy comparison with
10 results of PM25 source apportionment studies. Chow and Watson (2002) present a detailed
11 comparison of numerous studies using the CMB model performed mainly after 1995.
12 There are several differences between the broadly defined source categories shown at the
13 tops of Tables 3-9 and 3-10. These differences reflect the nature of sources that are important for
14 producing fine and coarse particulate matter shown in Table 3-8. They also are related to
15 improvements in the ability to distinguish between sources of similar nature (e.g., diesel and
16 gasoline vehicles, meat cooking, and vegetation burning). The use of organic tracers allows
17 motor vehicle emissions to be broken down into contributions from diesel and gasoline vehicles.
18 In studies where this distinction cannot be made, the source type is listed as 'total motor vehicles'
19 in the tables. The studies that were reported to be able to distinguish gasoline from diesel fueled
20 vehicles found that gasoline vehicles make significant, and sometimes the dominant,
21 contributions to ambient PM2 5 concentrations. Meat cooking is also distinguished from
22 vegetation burning in more recent studies, although both are considered to be part of biomass
23 burning. Vegetation burning consists of contributions from residential fuel wood burning,
24 wildfires, prescribed burning, and burning of agricultural and other biomass waste.
25 Miscellaneous sources of fine particles include contributions from combustion sources; whereas
26 miscellaneous sources of coarse particles consist of contributions from soil and sea spray and
27 industrial processing of geological material (e.g., cement manufacturing). Although a large
28 number of elements and chemical components are used to differentiate among source categories
29 and although there can be a large number of source types affecting a given site, only a few
30 broadly defined source types are needed to account for most of the mass of PM2 5 and PM10.
31
April 2002 3-76 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-9. RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
2 5
^
o
o
to
OJ
^
O
H
6
o
o
H
0
0
H
W
0
Sampling Site
Pasadena, CA 19821
Downtown LA, CA 19821
West LA, CA19821
Rubidoux, CA 19821
Sacramento, CA Winter 1991-962
Bakersfield, CA Winter 19963
Fresno, CA Winter 19963
Philadelphia, PA Summer 19824
Camden, NJ Summer 19824
Clarksboro, NJ Summer 19824
Grover City, IL ENEj 1986-875
Grover City, IL SSW 1986-875
Grover City, IL WNW 1986-875
Grover City, ILNNW 1986-875
Reno, NV Sumer 19986
Phoenix, AZ Summer 1995-987
Phoenix, AZ Winter 1995-987
'Schaueretal., 1996
2Motallebi, 1999
3Maglianoetal, 1998
"Dzubay etal, 1988
5Gloveretal, 1991
6Gilliesetal.,2000
'Ramadan etal., 2000
Total
Measured PM2 5 Motor Gasoline
Concentration Vehicles Diesel Vehicles
28.2 — 18.8 5.7
32.5 — 35.7 6.5
24.5 — 18.0 5.7
42.1 — 12.8 0.7
39.5 24.5e — —
52 16e — —
63 13e — —
27.0 8.5e — —
28.3 9.2e — —
26.0 5.8e — —
— — —
— — —
2.4e —
— — —
7.8 68e — —
8.3 — 10.9 36.2
13.8 — 14.5 38.9
Road Dust,
Soil
12.4
11.1
12.2
13.1
1.2
<3
<3
4.4
3.2
2.7
2.3
—
5.1
3.1
14.5
1.8
1.1
Secondary and other organic compounds
bSecondary ammonium
cMeat cooking
dVegetative detritus
eValue represents sum of diesel and gasoline
vehicle exhaust
Including associated cations and water
% Contribution3
Vegetation Secondary Secondary
Burning Sulfate Nitrate
9.6 20.9 7.4
5.8 20.3 9.2
11.0 24.1 7.8
1.2 13.8 24.7
18.1 4.5 36.6
20 7 34
19 5 32
— 81. 9f —
— 81.3f 0.4
— 84.6f —
— 83. 2f —
— 59.0f —
— 88. 5f —
— 86.6f —
4 11 2
15.0 — —
8.9 — —
incinerators
"Oil fly ash
Tluidized catalyst cracker
Wind direction
^Lead smelter
'Iron works
"Copper smelter
Misc. Misc. Misc. Misc.
Source 1 Source 2 Source 3 Source 4
5.3" 9.2b 8.5C l.ld
3.7" 9.2b 5.2° 0.6d
4.1" 9.4b 8.2° 1.6d
4.5" 12.1b 4.5C 0.5d
— — — —
— — — —
— — — —
2.28 1.9h 0.41 —
2.58 2.5h 0.71 —
0.88 1.5h 0.41 —
9.7k 3.01 1.2s —
11. 6k 11.91 4. 18 4.6m
2.8k — — —
3.41 3.0" — —
0.6" — — —
20.8" 4.9r 6.7s 3.6"
9.5" 4.5r 18.7s 4.1"
"Coal power plant
°As ammonium sulfate
pAs ammonium nitrate
qSea salt
Wood burning
TSTonferrous smelting
Total %
Allocated
98.9
107.3
102.1
87.9
84.9
<80
<85
99.3
99.8
95.8
99.4
91.2
98.8
96
100.1
99.9
100.2
O
-------
TABLE 3-10. RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
10
^
to
o
o
to
OJ
oo
o
£
H
1
O
o
o
H
O
o
H
W
0
^-i
% Contribution
Measured
PM10
Sampling Site Concentration
Craycroft, AZ
Winter 1 989-1 990y
Haydenl,AZ 1986Z
Hayden2,AZ 19861
Rillito,AZ 19882
Bakerfield, CA 1988-19893
Crows Landing, CA 1 988- 1 9893
Fellows, CA 1988-19893
Fresno, CA 1 988-1 9893
Indio, CA4
Kem Wildlife Refuge, CA
1988-19893
Long Beach, CA 19865
Long Beach, CA Summer 19876
Long Beach, CA Fall 19876
Riverside, CA 19887
Rubidoux, CA 19865
Rubidoux, CA Summer 19876
Rubidoux, CA Fall 19876
Rubidoux, CA 19887
San Nicolas Island, CA
Summer 19876
Stockton, CA 19893
Pocatello,ID 19908
S. Chicago, IL 19869
S.E. Chicago, IL 198810
Reno,NV 1986-87"
23.4
105.0
59.0
79.5
79.6
52.5
54.6
71.5
58.0
47.8
51.9
46.1
96.1
64.0
87.4
114.8
112.0
87.0
17.4
62.4
100.0
80.1
41.0
30.0
Primary
Geological
55.6
4.8
35.6
53.7
53.9
61.3
53.1
44.5
56.9
31.6
39.9
24.1
11.8
50.9
49.3
30.4
17.1
55.2
9.2
55.1
8.3
34.0
35. 9V
49.7
Primary
Motor
Primary Vehicle
Construction Exhaust
0.0
1.9b
6.8b
17.4b
2.0
0.0
2.6
0.0
5.2
4.2
0.0
0.0
0.0
0.0
4.6
3.9
14.4
0.0
0.0
0.8
7.5q
3.0
0.0
0.0
35.5
0.0
0.0
1.5f
9.7
4.2
3.8
9.5
7.6
4.6
9.81
13.7
44.5
10.9
6.41
15.1
27.1
11.7
5.2
8.3
0.1
3.5
2.2f
33.3
Primary
Vegetative
Burning
0.0
0.0
0.0
0.0
8.2
6.5
6.2
7.1
12.2
8.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
7.7
0.0
0.0
0.0
6.3
Secondary
Ammonium
Sulfate
3.0
3.8
6.8
0.0
6.9
5.3
9.3
5.0
6.2
6.9
15.4
23.6
4.0
7.5
7.3
8.3
1.9
6.1
21.3
5.0
0.0
19.2s
18.8
4.3
Secondary
Ammonium Misc.
Nitrate Source 1
2.6
0.0
0.0
0.0
16.0
12.4
13.7
14.5
7.1
3.1
17.7
1.7
24.1
33.4
24.4
23.9
28.2
24.9
2.9
11.2
0.0
—
2.0
5.1a
70.5C
47.5C
14.6g
1.3m
1.0m
12. 8m
0.4m
0.3j
1.0m
0.2]
0.2'
Off
0.5'
0.3]
0.0j
0.0]
0.6j
0.0>
l.lm
0.0
18.9'
2.0'
0.0
Misc.
Source 2
0.0
4.8d
0.0
0.0
1.9n
1.9n
2.6n
1.9n
1.7h
3.1n
3.9h
4.8h
2.8h
2.0h
l.lh
4.4h
1.0h
1.7h
24.7h
2.9n
0.0
2.T
0.7h
0.0
Misc.
Source 3
0.0
1.0e
1.7e
0.0
0.8k
2.3k
2.6k
O.lk
0.0
1.5k
12.3k
0.0
0.0
1.7°
6.8k
0.0
0.0
6.6°
0.0
0.0k
84. lr
0.0
2.7W
0.0
Misc.
Source 4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
18. 8g
0.0
Total %
Allocated
101.8
86.6
98.4
87.2
100.7
94.9
106.7
83
97.2
64.4
63.2
68.1
87.2
106.9
100.2
86
89.7
106.8
63.3
92.1
100
81.3
81.1
95.6
O
-------
TABLE 3-10 (cont'd). RECEPTOR MODEL SOURCE CONTRIBUTIONS TO PM
10
to
o
o
to
Sampling Site
Sparks, NV 1986-8711
Follansbee, WV 199 112
Mingo, OH 199112
Steubenville, OH 199 112
% Contribution
Measured
PM10
Concentration
41.0
66.0
60.0
46.0
Primary
Geological
36.8
15.2
20.0
18.0
Primary
Construction
0.0
0.0
0.0
0.0
Primary
Motor
Vehicle
Exhaust
28.3
53.0
23.3
30.4
Primary
Vegetative
Burning
32.7
0.0
6.8
1.7
Secondary
Ammonium
Sulfate
6.6
24.2
25.0
30.4
Secondary
Ammonium Misc.
Nitrate Source 1
2.2 0.0
— 14.1'
— 5.7'
— 8.3'
Misc.
Source 2
0.0
0.0
18. 3*
10.9*
Misc.
Source 3
0.5k
0.0
0.0
0.0
Misc.
Source 4
0.0
0.0
0.0
0.0
Total %
Allocated
107.1
106.5
99.1
99.7
OJ
VO
'Chowetal., 1992a
2Garfield; Ryan et al, 1988
3Jail; Ryan etal., 1988
"Thanukosetal., 1992
5Chowetal., 1992b
6Kim etal., 1992
'Gray etal., 1988
8Watsonetal., 1994
9Chowetal, 1992c
10Houck etal., 1992
"Hopke etal., 1988
12Vermetteetal., 1992
13Chow etal., 1988
"Skidmore et al., 1992
aSmelter background aerosol
bCement plant sources, including
kiln stacks, gypsum pile, and kiln
area
cCopper ore
"topper tailings
topper smelter building
'Heavy-duty diesel exhaust
emission
background aerosol
hMarine aerosol, road salt, and
sea salt plus sodium nitrate
'Motor vehicle exhaust from
diesel and leaded gasoline
^Residual oil combustion
Secondary organic carbon
'Biomass burning
""Primary crude oil
"NaCl + NaN03
"Lime
pRoad sanding material
qAsphalt industry
Thosphorus/phosphate industry
"Regional sulfate
'Steel mills
"Refuse incinerator
^Local road dust, coal yard road
dust, and steel haul road dust
"Incineration
^Unexplained mass
H
6
O
o
H
O
o
H
W
O
O
H
w
-------
1 At any given site, #5 source types account for >65% of the mass of PM2 5 (Table 3-9); and
2 #5 source types account for >65% of the mass of PM10 (Table 3-10).
3 Secondary sulfate is the dominant component of PM25 samples collected in the studies of
4 Dzubay et al. (1988) and Glover et al. (1991). Both studies found that sulfate at their monitoring
5 site arose from regionally dispersed sources. Sulfate also represents the major component of
6 PM25 found in monitoring studies in the eastern United States shown in Appendix 6 A of the
7 1996 PM AQCD. Primary and secondary organic components also make major contributions to
8 PM2 5. Contributions from road dust and soils are relatively minor, typically constituting less
9 than 10% of PM2 5 in the studies shown in Table 3-9. Studies in the western United States shown
10 in Table 3-9 have found larger contributions from motor vehicles, fugitive dust, and ammonium
11 nitrate. The most notable difference in the relative importance of major source categories of
12 PM25 shown in Table 3-9 and PM10 shown in Table 3-10 involves crustal material, (e.g., soil,
13 road dust), which represents about 40% on average of the total mass of PM10 in the studies shown
14 in Table 3-10. The fraction is higher at sites located away from specific sources such as sea
15 spray or smelters. Emissions of crustal material are concentrated mainly in the PM10_25 size
16 range.
17 In Table 3-10, primary motor vehicle exhaust contributions account for up to 40% of
18 average PM10 at many of the sampling sites. Vehicle exhaust contributions are also variable at
19 different sites within the same study area. The mean value and the variability of motor vehicle
20 exhaust contributions reflects the proximity of sampling sites to roadways and traffic conditions
21 during the time of sampling. Many studies were conducted during the late 1980s, when a portion
22 of the vehicle fleet still used leaded gasoline. Pb and Br in motor vehicle emissions facilitated
23 the distinction of motor vehicle contributions from other sources. Vehicles using leaded fuels
24 have higher emission rates than vehicles using unleaded fuels. Pb also poisons automobile
25 exhaust catalysts and produces adverse human health effects. As a result, Pb has been eliminated
26 from vehicle fuels. However, organic species such as n-pentacosane through n-nonacosene,
27 cholestanes, ergostanes, sitostanes, and hopanes have replaced Pb as a source marker for motor
28 vehicle emissions (e.g., Schauer and Cass, 2000). In their comprehensive review of CMB
29 modeling studies undertaken since 1995, Chow and Watson (2002) note that in twenty-two
30 studies fossil fuel combustion was found to be a large contributor to PM2 5 and PM10
April 2002 3-80 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations, with most of the contributions to primary PM originating from the exhaust of
2 diesel and gasoline vehicles.
3 Marine aerosol is found, as expected, at coastal sites such as Long Beach (average 3.8% of
4 total mass) and San Nicolas Island (25%). These contributions to PM10 are relatively variable
5 and are larger at the more remote sites. Individual values reflect proximity to local sources.
6 Of great importance are the contributions from secondary ammonium sulfate in the eastern
7 United States and ammonium nitrate in the western United States. Secondary ammonium sulfate
8 is especially noticeable at sites in California's San Joaquin Valley (Bakersfield, Crows Landing,
9 Fellows, Fresno, and Stockton) and in the Los Angeles area.
10 Because many source apportionment studies address problems in compliance with the
11 National Ambient Air Quality Standards and other air quality standards, samples selected for
12 chemical analysis are often biased toward the highest PM10 mass concentrations in the studies
13 shown in Table 3-10. Thus, the average source contribution estimates shown in Table 3-10 are
14 probably not representative of annual averages. For example, the study by Motallebi (1999)
15 considered only days when the PM10 concentration was greater than 40 //g/m3. Quoted
16 uncertainties in the estimated contributions of the individual sources shown in Tables 3-9 and
17 3-10 range from 10 to 50%. Errors can be much higher when the chemical source profiles for
18 different sources are highly uncertain or are too similar to distinguish one source from another.
19 Very few source apportionment studies using the CMB modeling technique have examined
20 the spatial variability of source contributions at different sites within an urban area. As can be
21 seen from Table 3-9, Dzubay et al. (1988) found a uniform distribution of sulfate among the NE
22 Airport in Philadelphia, PA; downtown Camden, NJ; and Clarksboro, NJ, during the summer of
23 1982. The farthest distance between two monitoring sites (NE Airport and Clarksboro) was
24 approximately 40 km. Magliano et al. (1998) examined the spatial variability of PM10 source
25 contributions at a number of sites in Fresno and Bakersfield, CA, during the winter of 1995-1996
26 and reported values for 1 day, December 27, 1995. During that day, mobile sources contributed
27 from 13.0 to!5.8 //g/m3, vegetation burning contributed from 5.1 to 11.1 Mg/m3, ammonium
28 sulfate contributed 2.4 to 3.4 //g/m3, and ammonium nitrate contributed 19.3 to 24.6 //g/m3 to
29 PM10 at the sites in Bakersfield. Mobile sources contributed 13.9 to 22.5 //g/m3, vegetation
30 burning contributed 8.2 to 15.7 //g/m3, ammonium sulfate contributed 1.8 to 2.3 //g/m3, and
31 ammonium nitrate contributed 14.5 to 18.9 //g/m3 at the sites in Fresno. All of these components
April 2002 3-81 DRAFT-DO NOT QUOTE OR CITE
-------
1 are expected to be found mainly in the PM2 5 size fraction. As can be seen, source contributions
2 at different sites varied by factors of 1.2 to 2.2 in Bakersfield and by factors of 1.3 to 1.9 in
3 Fresno on that day.
4 The receptor modeling methods outlined above do not explicitly include consideration of
5 the distances between PM sources and the receptor site. Information about the relative
6 importance of sources as a function of distance may be available from examination of data
7 obtained by continuous monitoring methods. For example, concentration spikes are expected to
8 be the result of transport from nearby sources, because turbulent mixing in the atmosphere would
9 not allow them to persist for very long. Short duration spikes in the time series of concentrations
10 are assumed to result from emissions from local sources (0.1 to 1 km away) in this method.
11 Contributions from sources located further away are determined by comparisons between
12 baselines measured at different sites. Details such as these are also lost in integrated 24-h
13 samples. Watson and Chow (2001) used time series of black carbon (BC) obtained by
14 aetholometers over five minute intervals to estimate the contributions from sources located
15 < 1 km away, 1 to 5 km away, and > 5 km away from a monitoring site in downtown Mexico
16 City. They found that most of the BC was produced by sources scattered throughout the city and
17 that sources located less than 1 km away from the site contributed only about 10% to BC
18 concentrations even in the presence of local sources such as buses and trucks.
19
20 3.3.4 Emissions Estimates for Primary Particulate Matter, and Precursors to
21 Secondary Particulate Matter (SO2, NOX, VOCs, and NH3) in the
22 United States
23 In principle, source contributions to ambient PM also could be estimated on the basis of
24 predictions made by chemistry-transport models (CTM) or even on the basis of emissions
25 inventories alone. Uncertainties in emissions inventories have arguably been regarded as
26 representing the largest source of uncertainty in CTMs (Calvert et al., 1993). Apart from
27 uncertainties in emission inventories, a number of other factors limit the ability of an emissions-
28 inventory driven CTM to determine the effects of various sources on particle samples obtained at
29 a particular location. CTM predictions represent averages over the area of a grid cell, which in
30 the case of CMAQ (Community Model for Air Quality) and MAQSIP (Multiscale Air Quality
31 Simulation Platform), ranges from 16 km2 (4 km x 4 km) to 1296 km2 (36 km x 36 km). CMAQ
April 2002 3-82 DRAFT-DO NOT QUOTE OR CITE
-------
1 and MAQSIP constitute the CTMs within the overall ModelsS framework, which also includes
2 emissions processors, the meteorological model, and modules for decision support. The
3 contributions of sources to pollutant concentrations at a monitoring site are controlled strongly by
4 local conditions that cannot be resolved by an Eulerian grid-cell model. Examples would be the
5 downward mixing of tall stack emissions and deviations from the mean flow caused by
6 buildings. The effect of local sources at a particular point in the model domain may not be
7 predicted accurately because their emissions would be smeared over the area of a grid cell or if
8 the local wind fields at the sampling point deviated significantly from the mean wind fields
9 calculated by the model. CTMs also have problems in predicting pollutant concentrations
10 because of uncertainties in vertical mixing and in predicting concentrations of pollutants from
11 stationary combustion sources resulting from uncertainties in estimates of plume rise. CTMs are
12 an integral part of air quality management programs and are reviewed in the NARSTO Fine
13 Particle Assessment (NARSTO, 2002).
14 Estimated emissions of primary PM25 from different sources in the United States are
15 summarized in Table 3-11, and estimated emissions of precursors to the formation of secondary
16 PM25 (SO2, NOX, VOCs, and NH3) are summarized in Table 3-12. These estimates are given to
17 provide a rough overview of the relative importance of major PM sources in the United States.
18 The emissions estimates are based on information presented in the EPA National Air Pollutant
19 Emission Trends Report, 1900-1999 (U.S. Environmental Protection Agency, 2001), to which
20 the reader is referred for detailed tables showing trends in PM2 5 emissions from a number of
21 source categories from 1990 to 1999. Detailed descriptions of the methodology for constructing
22 emissions inventories for criteria pollutants, quality assurance procedures, and examples of
23 calculations of emissions can be found in U. S. Environmental Protection Agency (1999).
24 Although uncertainties associated with the estimates in the National Air Pollutant Emission
25 Trends Report are not given therein, a discussion of uncertainties in emissions estimates is given
26 in Section 3.3.5.
27 For the sake of completeness, an attempt was made to supplement the information given in
28 the emissions tables in the Trends Report, which concentrates mainly on anthropogenic
29 emissions, with information about emissions from natural sources. Details regarding the
30 composition of the emissions of primary PM2 5 from the source categories shown in Table 3-11
31 are summarized in Appendix 3D, where available. Fugitive dust emissions are estimated to
April 2002 3-83 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-11. EMISSIONS OF PRIMARY PM?, BY VARIOUS SOURCES IN 1999
-2.5
Source
Emissions
(109 kg/y) Major PM Components
Notes
On-road vehicle 0.21
exhaust
Non-road vehicle 0.37
exhaust
Organic compounds,
elemental carbon
Organic compounds,
elemental carbon
Exhaust emissions from diesel (72%) and
gasoline vehicles (28%).
Exhaust emissions from off-road diesel (57%)
and gasoline vehicles (20%); ships and boats
(10%); aircraft (7%); railroads (6%).
Fossil fuel
combustion
Industrial
processes
Biomass burning
Waste disposal
Fugitive dust
Windblown dust
Other
Total
0.36 Crustal elements, trace
metals
0.35 Metals, crustal material,
organic compounds
1.2 Organic compounds,
elemental carbon
0.48 Organic compounds,
trace metals
3.3 Crustal elements
NA1 Crustal elements
0.02 Organic compounds,
elemental carbon
6.2
Fuel burning in stationary sources such as
power plants (33%); industries (39%);
businesses and institutions (25%); residences
(3%).
Metals processing (29%); mineral products
(27%); chemical mfg. (11%); other industries
(33%).
Managed burning (47%); residential wood
burning (28%); agricultural burning (7%);
wildfires (18%).
Open burning (91%); incineration (9%).
Dust raised by vehicles on paved (19%) and
unpaved roads (40%); construction (15%),
dust from raising crops (24%) and livestock
(2%).
Dust raised by wind on bare land.
Structural fires
'NA = not available.
Source: Adapted from U. S. Environmental Protection Agency (2001).
1 constitute over 50 percent of nationwide primary PM25 emissions, according to Table 3-11.
2 However, there are a number of issues concerning the methods for obtaining relevant emissions
3 factor data for fugitive dust in field studies, as discussed in Section 3.3.5. An estimate of the
4 production of PM25 from wind erosion on natural surfaces was not included in Table 3-11
5 because this source is highly sporadic, occurs during periods of high winds and, thus, the
6 resulting emissions are too highly uncertain to be included. As can be seen from a comparison of
7 entries in Tables 3-11 and 3-12, estimates of emissions of potential precursors to secondary PM
8 formation are considerably larger than those for estimates of primary PM25 emissions in the
April 2002
3-84
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3-12. EMISSIONS OF PRECURSORS TO SECONDARY PM2 5 FORMATION
BY VARIOUS SOURCES IN 1999
Precursor
Emissions
(109kg/y)
Secondary PM
Component
Notes
SO,
NCL
17 Sulfate
26 Nitrate
Anthropogenic 16
VOCs
Various mainly
unidentified
compounds of 'OC'
Biogenic
VOCs1
NH,
44 Various mainly
unidentified
compounds of 'OC'
45 Ammonium
Exhaust from on-road (2%) and non-road (5%) engines
and vehicles; fossil fuel combustion by electrical utilities,
industries, other sources (85%); various industrial
processes (7%); and other minor sources (1%).
Exhaust from on-road (34%) and non-road (22%) engines
and vehicles; fossil fuel combustion by electrical utilities,
industries, other sources (39%); lightning (4%); soils
(4%); and other minor sources (5%).
Evaporative and exhaust emissions from on-road (29%)
and non-road (18%) vehicles; evaporation of solvents and
surface coatings (27%); biomass burning (9%); storage
and transport of petroleum and volatile compounds (7%);
chemical and petroleum industrial processes (5%); other
sources (5%).
Approximately 98% emitted by vegetation. Isoprene
(35%); monoterpenes (25%); all other reactive and
non-reactive compounds (40%).
Exhaust from on-road and non-road engines and vehicles
(5%); chemical manufacturing (3%); waste disposal,
recycling, and other minor sources (5%); livestock (82%);
and fertilizer application (18%).
'Includes estimates of natural sources from Guenther et al. (2000).
Emissions expressed in terms of NO2.
Source: Adapted from U. S. Environmental Protection Agency (2001).
1 United States. The emissions of SO2, NOX, and NH3 should be multiplied by factors of 1.5, 1.35,
2 and 1.07, respectively, to account for their chemical form in the aerosol phase. Estimating a
3 factor for VOCs is somewhat less straight forward. Turpin and Lim (2001) recommends a factor
4 of 2 to account for the conversion of VOC precursors to oxygen and nitrogen containing
5 compounds in the aerosol phase. These factors are all greater than 1 and further underscore the
6 potential importance of secondary PM precursor emissions relative to primary PM emissions.
7 However, the emissions of precursors cannot be translated directly into rates of PM formation.
8 Dry deposition and precipitation scavenging of some of these gaseous precursors and their
9 intermediate oxidation products occur before they are converted to PM in the atmosphere.
April 2002
3-85
DRAFT-DO NOT QUOTE OR CITE
-------
1 In addition, some fraction of these gases are transported outside of the domain of the continental
2 United States before being oxidized. Likewise, emissions of these gases from areas outside the
3 United States can result in the transport of their oxidation products into the United States.
4 As discussed in Section 3.3.1, the photochemical oxidation of sulfur dioxide leads to the
5 production of sulfate; whereas that of nitrogen oxide leads ultimately to particulate-phase nitrite
6 and nitrate. Due to uncertainties it is difficult to calculate the rates of formation of secondary
7 organic particulate matter (SOPM) from the emissions of VOC precursors. Smog chamber and
8 laboratory studies discussed in Section 3.3.1 indicate that anthropogenic aromatic compounds
9 and biogenic terpenoid compounds have the highest potential for forming secondary organic
10 parti culate matter; and as can be seen from Table 3C-1, the dominant compounds tend to be
11 those derived from these categories. Each of the source categories capable of emitting VOCs
12 shown in Table 3-12 has components capable of forming SOPM, although in small yields
13 (ranging typically up to several per cent, cf Section 3.3.1). The oxidation of lighter organic
14 compounds leads ultimately to the formation of CO and CO2. As discussed by Pandis et al.
15 (1991) and in Section 3.3.1, soluble gas phase compounds, such as formaldehyde (CH2O), other
16 aldehydes, organic acids, etc. formed during the oxidation of a wide variety of hydrocarbons, can
17 be incorporated into suspended particles. Although isoprene is a major component of biogenic
18 emissions, its oxidation has not been found to result in the formation of new particles; whereas
19 the oxidation of monoterpenes has. However, it should be remembered that soluble gas phase
20 species such as CH2O are formed during the oxidation of isoprene.
21 The emissions estimates shown in this section are based on annual totals. However, annual
22 averages do not reflect the variability of a number of emissions categories on shorter time scales.
23 Residential wood burning in fireplaces and stoves, for example, is a seasonal practice that
24 reaches its peak during cold weather. Cold weather also affects motor vehicle exhaust particulate
25 matter emissions, both in terms of chemical composition and emission rates (e.g., Watson et al.,
26 1990b; Huang et al., 1994). Agricultural activities such as planting, fertilizing, and harvesting
27 are also seasonal. Forest fires occur mainly during the local dry season and during periods of
28 drought. Maximum dust production by wind erosion in the United States occurs during the
29 spring; whereas the minimum occurs during the summer (Gillette and Hanson, 1989). Efforts are
30 being made to account for the seasonal variations of emissions in the nationwide emissions
April 2002 3-86 DRAFT-DO NOT QUOTE OR CITE
-------
1 inventories. Techniques for calculating emissions of criteria pollutants on a seasonal basis are
2 given in U. S. Environmental Protection Agency (1999).
3 Trends in nationwide, annual average concentrations of PM10, and precursor gases (SO2,
4 NO2, and VOC) over the 10 years from 1989 to 1998 are shown in Table 3-13. As can be seen
5 from Table 3-13, there have been substantial decreases in the ambient concentrations of PM10,
6 SO2, and NO2. Not enough data are available to define trends in concentrations of VOCs. There
7 also have been substantial decreases in the emissions of all the species shown in Table 3-13,
8 except for NO2, although its average ambient concentration has decreased by 14%. These entries
9 suggest that decreases in the average ambient concentration of PM10 could have been produced
10 by both decreases in emissions of primary PM10 and the formation of secondary PM10. The large
1 1 reductions in ambient SO2 concentrations have resulted in reductions in sulfate formation that
12 would have been manifest in PM2 5 concentrations on the regional scale in the East and Midwest,
13 where sulfate has constituted a larger fraction of PM2 5 than in the West. Likewise, reductions in
14 NO2 concentrations would have had a more noticeable effect on PM2 5 concentrations in the West
15 than in the East, because nitrate is a larger component of the aerosol in the West.
16
17
TABLE 3-13. NATIONWIDE CHANGES IN AMBIENT CONCENTRATIONS AND
EMISSIONS OF PM10 AND GASEOUS PRECURSORS TO SECONDARY
PARTICULATE MATTER FROM 1990 TO 1999
% Change 1990-1999
PM10
PM2,
(1992 to 1999)
SO47SO2
NO3-/NOX
VOC
Ambient Concentration
-18%
Urban east -2%
Rural east - 5
Rural west -15%
-3 6% (sulfate)
-10% (nitrate)
—
Emissions
-15%
-17%
-20% (SO2)
+5% (NOX)
-14%
Source: U. S. Environmental Protection Agency (2000d).
April 2002 3-87 DRAFT-DO NOT QUOTE OR CITE
-------
1 Trends in aerosol components (i.e., nitrate, sulfate, carbon, etc.) are needed for a more
2 quantitative assessment of the effects of changes in emissions of precursors. Aerosol nitrate and
3 sulfate concentrations obtained at North Long Beach and Riverside, CA, tracked downward
4 trends in NOX concentrations. SO2 and sulfate concentrations have both decreased; however, the
5 rate of decline of sulfate has been smaller than that of SO2, indicating the long range transport of
6 sulfate from outside the air shed may be an important source in addition to the oxidation of
7 locally generated SO2. There are a number of reasons why pollutant concentrations do not track
8 estimated reductions in emissions. Some of these reasons are related to atmospheric effects such
9 as meteorological variability and secular changes in the rates of photochemical transformations
10 and deposition (U.S. Environmental Protection Agency, 2000c). Other reasons are related to
11 uncertainties in ambient measurements and in emissions inventories.
12
13 3.3.5 Uncertainties of Emissions Inventories
14 As described in the 1996 PM AQCD, it is difficult to assign uncertainties quantitatively to
15 entries in emissions inventories. Methods that can be used to verify or place constraints on
16 emissions inventories are sparse. In general, the overall uncertainty in the emissions of a given
17 pollutant includes contributions from all of the terms used to calculate emissions (i.e., activity
18 rates, emissions factors, and control device efficiencies). Additional uncertainties arise during
19 the compilation of an emissions inventory because of missing sources and computational errors.
20 The variability of emissions can cause errors when annual average emissions are applied to
21 applications involving shorter time scales.
22 Activity rates for well-defined point sources (e.g., power plants) should have the smallest
23 uncertainty associated with their use because emissions are monitored continuously in many
24 cases accurate production records need to be kept. On the other hand, activity rates for a number
25 of very dispersed fugitive sources are difficult to quantify. Emissions factors for easily measured
26 fuel components that are released quantitatively during combustion (e.g., CO2, SO2) should be
27 the most reliable. Emissions of components formed during combustion are more difficult to
28 characterize, as the emissions rates are dependent on factors specific to individual combustion
29 units and on combustion stage (i.e., smoldering or active). Although the AP-42 emissions factors
30 (U.S. Environmental Protection Agency, 1995) contain extensive information for a large number
31 of source types, these data are very limited in the number of sources sampled. The efficiency of
April 2002 3-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 control devices is determined by their design, their age, their maintenance history, and operating
2 conditions. It is virtually impossible to assign uncertainties in control device performance
3 because of these factors. It should be noted that the largest uncertainties occur for those devices
4 that have the highest efficiencies (>90%). This occurs because the efficiencies are subtracted
5 from one, and small errors in assigning efficiencies can lead to large errors in emissions.
6 Ideally, an emissions inventory should include all major sources of a given pollutant. This
7 may be an easy task for major point sources. However, area sources of both primary PM and
8 precursors to secondary PM formation are more difficult to characterize than point sources; and,
9 thus, they require special emphasis when preparing emission inventories. Further research is
10 needed to better characterize the sources of pollutants to reduce this source of uncertainty. Errors
11 also can arise from the misreporting of data, and arithmetic errors can occur in the course of
12 compiling entries from thousands of individual sources. A quality assurance program is required
13 to check for outliers and arithmetic errors.
14 Because of the variability in emissions rates, there can be errors in the application of
15 inventories developed on an annually averaged basis (as are the inventories shown in Tables 3-11
16 and 3-12) to episodes occurring on much shorter time scales. As an example, most modeling
17 studies of air pollution episodes are carried out for periods of a few days.
18 Uncertainties in annual emissions were estimated to range from 4 to 9% for SO2 and from
19 6 to 11% for NOX in the 1985 NAPAP inventories for the United States (Placet et al., 1991).
20 Uncertainties in these estimates increase as the emissions are disaggregated both spatially and
21 temporally. The uncertainties quoted above are minimum estimates and refer only to random
22 variability about the mean assuming that the variability in emissions factors was adequately
23 characterized and that extrapolation of emissions factors to sources other than those for which
24 they were measured is valid. The estimates do not consider the effects of weather or variations in
25 operating and maintenance procedures.
26 Fugitive dust sources, as mentioned above, are extremely difficult to quantify; and stated
27 emission rates may represent only order-of-magnitude estimates. Although crustal dust
28 emissions constitute about 50% of the total primary PM25 inventory, they constitute less than
29 about 15% of the source strengths inferred from the receptor modeling studies shown in
30 Table 3-9. However, it should be remembered that secondary components (sulfate, nitrate, and
31 some fractions of organic carbon) often account for most of the mass of ambient PM2 5 samples.
April 2002 3-89 DRAFT-DO NOT QUOTE OR CITE
-------
1 Although mineral dust sources represent the major category in Table 3-11, their
2 contributions are distributed much more widely than are those from combustion sources. Watson
3 and Chow (1999) reexamined the methodology used to determine emissions of fugitive dust.
4 The standard methods use data obtained by particle monitors stacked at several elevations from
5 1 to 2 m up to 7 to 10 m above the surface. However, small-scale turbulent motions and variable
6 winds characterize atmospheric flow patterns immediately adjacent to the surface (Garratt, 1994).
7 The depth of this turbulent layer is determined by surface roughness elements; and, if particle
8 monitors are sampling within this layer, there is a high probability of particles being entrained in
9 turbulent eddies and redepositing on the ground within a very short distance. In addition to the
10 source sampling problem referred to above, it should be remembered that dust often is raised in
11 remote areas far removed from population centers. Precipitation or scavenging by cloud droplets
12 and dry deposition removes particles during transport from the source area. In addition,
13 gravitational settling can be an important loss mechanism for particles larger than a few
14 micrometers in aerodynamic diameter.
15 As rough estimates, uncertainties in emissions estimates could be as low as 10% for the
16 best characterized source categories; whereas emissions figures for windblown dust should be
17 regarded as order-of-magnitude estimates. The application of emissions inventories to the
18 estimation of source contributions at monitoring sites is also limited by the effects of local
19 topography and meteorology. For example, Pinto et al. (1998) found that the contribution of
20 power plants and residential space heating to PM2 5 concentrations in northwestern Bohemia are
21 comparable on the basis of CMB receptor modeling. However, according to the emissions
22 inventories, the contribution from power plants should have been roughly an order of magnitude
23 larger than that from residential space heating. The difference between the two methods can be
24 explained by noting that mixing of the emissions from the power plants downward to the surface
25 is inhibited by strong surface inversions that develop during the winter season in this area.
26 There have been few field studies designed to test emissions inventories observationally.
27 The most direct approach would be to use aircraft to obtain cross-sections of pollutants upwind
28 and downwind of major urban areas. The computed mass flux through a cross section of the
29 urban plume can then be equated to emissions from the city chosen. This approach has been
30 attempted on a few occasions, but results have been ambiguous because of contributions from
31 fugitive sources, variable wind flows, and logistic difficulties.
April 2002 3-90 DRAFT-DO NOT QUOTE OR CITE
-------
1 3.4 SUMMARY AND CONCLUSIONS
2 The recently deployed PM2 5 FRM network has returned data for a large number of sites
3 across the United States. Annual mean PM2 5 concentrations range from about 5 //g/m3 to about
4 30 //g/m3. In the eastern United States, the data from 1999 and 2000 indicate that highest
5 quarterly mean concentrations and maximum concentrations occur during the summer. In the
6 western United States, highest quarterly mean values and maximum values occur mainly during
7 the winter at a number of sites, although there were exceptions to these general patterns. These
8 findings are generally consistent with those based on longer term data sets such as MAAQS in
9 the eastern United States and the CARB network of dichotomous samplers in California. PM2 5
10 and PM10 concentrations in a number of urban areas have generally declined over the past few
11 decades. However, they have leveled off in the past few years.
12 Differences in annual mean PM25 concentrations between monitoring sites in urban areas
13 examined are typically less than 6 or 7 //g/m3. However, on individual days, differences in 24-h
14 average PM2 5 concentrations can be much larger. Some sites in metropolitan areas are highly
15 correlated with each other, but other sites are not due to the presence of local sources,
16 topographic barriers, etc. Although PM25 concentrations at sites within an MSA can be highly
17 correlated, there still can be significant differences in their concentrations on any given day.
18 Consequently, additional measures should be used to characterize the spatial variability of PM25
19 concentrations. The degree of spatial uniformity in PM2 5 concentrations in urban areas varies
20 across the country. These factors should be considered in using data obtained by the PM2 5 FRM
21 network to approximate community-scale human exposure, and caution should be exercised in
22 extrapolating conclusions obtained in one urban area to another. PM2 5 to PM10 ratios were
23 generally higher in the East than in the West, and values for this ratio are consistent with those
24 found in numerous earlier studies presented in the 1996 PM AQCD.
25 Data for PM10_2 5 are not as abundant as they are for PM2 5, and their interpretation is
26 complicated by the difference method used to determine their concentrations. The more sporadic
27 nature of sources of PM10_2 5 and its shorter atmospheric lifetime tend to result in lower
28 correlations for PM10_2 5 than for PM2 5 concentrations. Errors in measurement of PM2 5 and PM10
29 also result in lower spatial correlations of PM10_25. Calculated concentrations of PM10_25 are
30 occasionally negative as reflected by PM2 5 to PM10 ratios greater than one. Because analytical
31 errors are generally larger for individual species than for total mass, similar problems arise in
April 2002 3-91 DRAFT-DO NOT QUOTE OR CITE
-------
1 their determination in PM10_25 samples by the difference approach. Some, but not all of these
2 problems could be resolved by the use of dichotomous samplers that also provide a direct sample
3 of PM10_25 for compositional analyses.
4 Estimates of concentrations of individual species in PM10_25 samples were limited to those
5 obtained by dichotomous samplers. Generally, concentrations of most elements differ for PM2 5
6 and PM10_25. However, the available data suggest that concentrations of many metals are of the
7 same order of magnitude in both size fractions. This is in marked contrast to the situation twenty
8 years ago, when uncontrolled combustion sources were prevalent. At that time, concentrations of
9 many metals, especially lead, were much higher than today in fine-mode particles, and their
10 concentrations were much higher in the fine-mode than in the coarse-mode. No substantive
11 conclusions about contemporary concentrations and composition of ultrafine particles
12 (0.1 //m < Da) can be drawn for the nation as a whole, because of a lack of data.
13 Ambient PM contains both primary and secondary components. The results of ambient
14 monitoring studies and receptor modeling studies indicate that PM2 5 is dominated by secondary
15 components in the eastern United States. Depending on the origin of OC in ambient samples,
16 PM25, on average, may also be dominated by secondary components throughout the rest of the
17 United States. Primary constituents represent smaller but still important components of PM25.
18 Crustal materials, which are primary constituents, constitute the largest measured fraction of
19 PM10_2 5 throughout the United States. Data for the concentration of bioaerosols in both the PM2 5
20 and PM10_2 5 size ranges are sparse. Data collected in several airsheds, including the Los Angeles
21 Basin, Bakersfield and Fresno, CA; and Philadelphia, PA, suggest that secondary PM
22 components are more uniformly distributed than are primary PM components. Compositional
23 data obtained at multiple sites in other urban areas are sparse.
24 Because of the complexity of the composition of ambient PM25 and PM10_25, sources are
25 best discussed in terms of individual constituents of both primary and secondary PM25 and
26 PM10_25. Each of these constituents can have anthropogenic and natural sources, as shown in
27 Table 3-8. The distinction between natural and anthropogenic sources is not always obvious.
28 Although windblown dust might seem to be the result of natural processes, highest emission rates
29 are associated with agricultural activities in areas that are susceptible to periodic drought.
30 Examples include the dust bowl region of the midwestern United States and the Sahel of Africa.
31 There is also ongoing debate about characterizing wild fires as either natural or anthropogenic.
April 2002 3-92 DRAFT-DO NOT QUOTE OR CITE
-------
1 Land management practices and other human actions affect the occurrence and scope of wild
2 fires. Similarly, prescribed burning can be viewed as anthropogenic, or as a substitute for wild
3 fires that would otherwise occur eventually on the same land.
4 Over the past decade, a significant amount of research has been carried out to improve the
5 understanding of the atmospheric chemistry of secondary organic PM formation. Although
6 additional sources of SOPM might still be identified, there appears to be a general consensus that
7 biogenic compounds (monoterpenes, sesquiterpenes) and aromatic compounds (toluene,
8 ethylbenzene) are the most significant SOPM precursors. A large number of compounds have
9 been detected in biogenic and aromatic SOPM, although the chemical composition of these two
10 categories has not been fully established, especially for aromatic SOPM. Transformations that
11 occur during the aging of particles are still inadequately understood. There are still large gaps in
12 the current understanding of a number of key processes related to the partitioning of semivolatile
13 compounds between the gas phase and ambient particles containing organic compounds, liquid
14 water, and inorganic salts and acids. In addition, there is a general lack of reliable analytical
15 methods for measuring multifunctional oxygenates in the gas and aerosol phases.
16 The results of receptor modeling studies throughout the United States indicate that the
17 combustion of fossil and biomass fuels is the major source of measured ambient PM2 5. Fugitive
18 dust, found mainly in the PM10_25 range size, represents the largest source of measured ambient
19 PM10 in many locations in the western United States. Quoted uncertainties in the source
20 apportionment of constituents in ambient aerosol samples typically range from 10 to 50%. It is
21 apparent that a relatively small number of source categories, compared to the total number of
22 chemical species that typically are measured in ambient monitoring-source receptor model
23 studies, are needed to account for the majority of the observed mass of PM in these studies.
24 As seen in Table 3-8, emissions of mineral dust, organic debris, and sea spray are
25 concentrated mainly in the coarse fraction of PM10 (>2.5 //m aerodynamic diameter). A small
26 fraction of this material is in the PM2 5 size range (< 2.5 //m aerodynamic diameter). Still, PM2 5
27 concentrations of crustal material can be appreciable, especially during dust events. It also
28 should be remembered that from one-third to one-half of the Saharan dust reaching the United
29 States is in the PM2 5 size range. Emissions from combustion sources (mobile and stationary
30 sources and biomass burning) are also predominantly in the PM2 5 size range.
April 2002 3-93 DRAFT-DO NOT QUOTE OR CITE
-------
1 Although most emphasis in this chapter has been on sources within the United States,
2 it should be remembered that sources outside the United States contribute to ambient PM levels
3 that can, at times, exceed the ambient NAAQS level for PM. Dust is frequently transported from
4 northern Africa to the eastern United States. This dust often produces dense haze during the
5 summer in southern Florida. Bioaerosols and pollutants are also transported with the dust.
6 Large-scale dust storms in the deserts of central Asia recently have been found to contribute to
7 PM levels in the northwestern United States on an episodic basis. Uncontrolled biomass burning
8 in central America and Mexico may have contributed to elevated PM levels that exceeded the
9 daily NAAQS level for PM in Texas. Wildfires throughout the United States, Canada, Mexico,
10 and Central America all contribute to background concentrations of PM in the United States.
11
April 2002 3-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Altshuller, A. P. (1983) Review: natural volatile organic substances and their effect on air quality in the United
3 States. Atmos. Environ. 17: 2131-2165.
4 Ansari, A. S.; Pandis, S. N. (1998) Response of inorganic PM to precursor concentrations. Environ. Sci. Technol.
5 32:2706-2714.
6 Atkinson, R. (1994) Gas-phase tropospheric chemistry of organic compounds. Washington, DC: American
7 Chemical Society. (J. Phys. Chem. Ref. Data Monograph no. 2).
8 Aumont, B.; Madronich, S.; Bey, I.; Tyndall, G. S. (2000) Contribution of secondary VOC to the composition of
9 aqueous atmospheric particles: a modeling approach. J. Atmos. Chem. 35: 59-75.
10 Bahadori, T.; Suh, H.; Koutrakis, P. (2000a) A cross-study analysis of ambient concentrations of fine mass. J. Air
11 Waste Manage. Assoc.: in press.
12 Bahadori, T.; Suh, H.; Koutrakis, P. (2000b) PM2 5 and PM10 concentrations in four metropolitan cities in eastern
13 United States. J. Air Waste Manage. Assoc.: in press.
14 Cahill, T. A. (1996) [Data from IMPROVE data base.] Fort Collins, CO: Air Resource Specialists, Inc.
15 California Air Resources Board. (1995) [Data base provided by John R. Holmes, Chief, Research Division.]
16 Sacramento, CA: California Air Resources Board.
17 Calvert, J. G.; Heywood, J. B.; Sawyer, R. F.; Seinfeld, J. H. (1993) Achieving acceptable air quality: some
18 reflections on controlling vehicle emissions. Science (Washington, DC) 261: 37-45.
19 Cass, G. R. ; Hughes, L. A.; Bhave, P.; Kleeman, M. J.; Allen, J. O.; Salmon, L. G. (2000) The chemical
20 composition of atmospheric ultrafine particles. Philos. Trans. R. Soc. London Ser. A 358: 2581-2592.
21 Castro, L. M.; Pio, C. A.; Harrison, R. M.; Smith, D. J. T. (1999) Carbonaceous aerosol in urban and rural
22 European atmospheres: estimation of secondary organic carbon concentrations. Atmos. Environ. 33:
23 2771-2781.
24 Chadwick, O. A.; Deny, L. A.; Vitousek, P. M.; Huebert, B. J.; Hedin, L. O. (1999) Changing sources of nutrients
25 during four million years of ecosystem development. Nature (London, UK) 397: 491-497.
26 Chow, J. C.; Watson, J. G. (2002) Review of PM2 5 and PM10 apportionment for fossil fuel combustion and other
27 sources by the chemical mass balance receptor model. Energy Fuels: in press.
28 Chow, J. C.; Watson, J. G.; Frazier, C. A.; Egami, R. T.; Goodrich, A.; Ralph, C. (1988) Spatial and temporal
29 source contributions to PM10 and PM2 5 in Reno, NV. In: Mathai, C. V.; Stonefield, D. H., eds. PM.10:
30 implementation of standards, an APCA/EPA international specialty conference; February; San Francisco,
31 CA. Pittsburgh, PA: Air Pollution Control Association; pp. 438-457. (APCA transactions series TR-13).
32 Chow, J. C.; Watson, J. G.; Richards, L. W.; Haase, D. L.; McDade, C.; Dietrich, D. L.; Moon, D.; Sloane, C.
33 (1991) The 1989-90 Phoenix PM10 study. Volume II: source apportionment. Final report. Phoenix, AZ:
34 Arizona Department of Environmental Quality; Desert Research Institute document no. 8931.6F1.
35 Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Frazier, C. A.; Hinsvark, B. A.; Prichett, L. C.; Neuroth, G. R.
36 (1992a) Wintertime PM10 and PM2 5 chemical compositions and source contributions in Tucson, Arizona.
37 In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate source controls, an
38 A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air &
39 Waste Management Association; pp. 231-243. (A&WMA transactions series no. 22).
40 Chow, J. C.; Watson, J. G.; Lowenthal, D. H.; Solomon, P. A.; Magliano, K. L.; Ziman, S. D.; Richards, L. W.
41 (1992b) PM10 source apportionment in California's San Joaquin Valley. Atmos. Environ. Part A
42 26: 3335-3354.
43 Chow, J. C.; Liu, C. S.; Cassmassi, J.; Watson, J. G.; Lu, Z.; Pritchett, L. C. (1992c) A neighborhood-scale study of
44 PM10 source contributions in Rubidoux, California. Atmos. Environ. Part A 26: 693-706.
45 Chow, J. C.; Watson, J. G.; Edgerton, S. A.; Vega, E.; Ortiz, E. (2001) Spatial differences in outdoor PM10 mass
46 and aerosol composition in Mexico City. J. Air Waste Manage. Assoc.: submitted.
47 Christoffersen, T. S.; Hjorth, J.; Horie, O.; Jensen, N. R.; Kotzias, D.; Molander, L. L.; Neeb, P.; Ruppert, L.;
48 Winterhalter, R.; Virkkula, A.; Wirtz, K.; Larsen, B. R. (1998) cis-Pinic acid, a possible precursor for organic
49 aerosol formation from ozonolysis of a-pinene. Atmos. Environ. 32: 1657-1661.
50 Chung, A.; Herner, J. D.; Kleeman, M. J. (2001) Detection of alkaline ultrafine atmospheric particles at Bakersfield,
51 California. Environ. Sci. Technol. 35: 2184-2190.
52 Cocker, D. R., Ill; Clegg, S. L.; Flagan, R. C.; Seinfeld, J. H. (2001) The effect of water on gas-particle partitioning
53 of secondary organic aerosol. Part I: a-pinene/ozone system. Atmos. Environ. 35: 6049-6072.
April 2002 3-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 Darlington, T. L.; Kahlbaum, D. F.; Heuss, J. M.; Wolff, G. T. (1997) Analysis of PM10 trends in the United States
2 from 1988 through 1995. J. Air Waste Manage. Assoc. 47: 1070-1078.
3 Davis, A. P.; Shokouhian, M.; Ni, S. (2001) Loading estimates of lead, copper, cadmium, and zinc in urban runoff
4 from specific sources. Chemosphere 44: 997-1009.
5 Dolislager, L. J.; Motallebi, N. (1999) Characterization of paniculate matter in California. J. Air Waste Manage.
6 Assoc. 49: PM-45-56.
7 Duce, R. A. (1995) Sources, distributions, and fluxes of mineral aerosols and their relationship to climate.
8 In: Charlson, R. J.; Heintzenberg, J., eds. Aerosol forcing of climate: report of the Dahlem workshop on
9 aerosol forcing of climate; April 1994; Berlin, Federal Republic of Germany. Chichester, United Kingdom:
10 John Wiley & Sons, Ltd.; pp. 43-72.
11 Dzubay, T. G.; Stevens, R. K.; Gordon, G. E.; Olmez, I.; Sheffield, A. E.; Courtney, W. J. (1988) A composite
12 receptor method applied to Philadelphia aerosol. Environ. Sci. Technol. 22: 46-52.
13 Edney, E. O.; Driscoll, D. J.; Weathers, W. S.; Kleindienst, T. E.; Conver, T. S.; Mclver, C. D.; Li, W. (2001)
14 Formation of polyketones in irradiated toluene/propylene/NOx/air mixtures. Aerosol Sci. Technol.
15 35:998-1008.
16 Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the
17 United States of America. Atmos. Environ. 28: 1009-1019.
18 Faust, B. C. (1994) Photochemistry of clouds, fogs, and aerosols. Environ. Sci. Technol. 28: 217A-222A.
19 Fitz-Simons, T. S.;Mathias, S.;Rizzo,M. (2000) Analyses of 1999 PM data for the PMNAAQS review. Research
20 Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
21 November 17. Available: http://www.epa.gov/oar/oaqps/pm25/analyses.html [2 April, 2002].
22 Forstner, H. J. L.; Flagan, R. C.; Seinfeld, J. H. (1997a) Molecular speciation of secondary organic aerosol from
23 photooxidation of the higher alkenes: 1-octene and 1-decene. Atmos. Environ. 31: 1953-1964.
24 Forstner, H. J. L.; Flagan, R. C.; Seinfeld, J. H. (1997b) Secondary organic aerosol from the photooxidation of
25 aromatic hydrocarbons: molecular composition. Environ. Sci. Technol. 31: 1345-1358.
26 Garratt, J. R. (1994) The atmospheric boundary layer. Cambridge, United Kingdom: Cambridge University Press.
27 (Cambridge atmospheric and space science series).
28 Gatz, D. F.; Prospero, J. M. (1996) A large silicon-aluminum aerosol plume in central Illinois: North African Desert
29 dust? Atmos. Environ. 30: 3789-3799.
30 Gillette, D. A.; Hanson, K. J. (1989) Spatial and temporal variability of dust production caused by wind erosion in
31 the United States. J. Geophys. Res. [Atmos.] 94: 2197-2206.
32 Gillies, J. A.; Gertler, A. W.; Lowenthal, D.; Abu-Allaban, M.; Jennison, B.; Goodrich, A. (2000) Enhanced PM25
3 3 source apportionment using chemical mass balance receptor modeling and scanning electron microscopy.
34 Aerosol Sci. Technol.: submitted.
35 Glasius, M.; Lahaniati, M.; Calogirou, A.; Di Bella, D.; Jensen, N. R.; Hjorth, J.; Kotzias, D.; Larsen, B. R. (2000)
36 Carboxylic acids in secondary aerosols from oxidation of cyclic monoterpenes by ozone. Environ. Sci.
37 Technol. 34: 1001-1010.
38 Glover, D. M.; Hopke, P. K.; Vermette, S. J.; Landsberger, S.; D'Auben, D. R. (1991) Source apportionment with
39 site specific source profiles. J. Air Waste Manage. Assoc. 41: 294-305.
40 Gone, J.-K.; Olmez, I.; Ames, M. R. (2000) Size distribution and probable sources of trace elements in submicron
41 atmospheric paniculate material. J. Radioanal. Nucl. Chem. 244: 133-139.
42 Graedel, T. E.; Goldberg, K. I. (1983) Kinetic studies of raindrop chemistry: 1. Inorganic and organic processes.
43 J. Geophys. Res. C: Oceans Atmos. 88: 10,865-10,882.
44 Gray, H. A.; Landry, B.; Liu, C. S.; Henry, R. C.; Cooper, J. A.; Sherman, J. R. (1988) Receptor modeling for PM10
45 source apportionment in the South Coast Air Basin of California. In: Mathai, C. V.; Stonefield, D. H., eds.
46 PM.10: implementation of standards, an APCA/EPA international specialty conference; February;
47 San Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 399-418. (APCA transactions
48 series TR-13).
49 Griffin, R. J.; Cocker, D. R., Ill; Seinfeld, J. H. (1999) Incremental aerosol reactivity: application to aromatic and
50 biogenic hydrocarbons. Environ. Sci. Technol. 33: 2403-2408.
51 Griffin, D. W.; Garrison, V. H.; Herman, J. R.; Shinn, E. A. (2001) African desert dust in the Caribbean
52 atmosphere: microbiology and public health. Aerobiologia 17: 203-213.
53 Grosjean, D. (1992) In situ organic aerosol formation during a smog episode: estimated production and chemical
54 functionality. Atmos. Environ. Part A 26: 953-963.
April 2002 3-96 DRAFT-DO NOT QUOTE OR CITE
-------
1 Grosjean, D.; Seinfeld, J. H. (1989) Parameterization of the formation potential of secondary organic aerosols.
2 Atmos. Environ. 23: 1733-1747.
3 Guenther, A.; Geron, C.; Pierce, T.; Lamb, B.; Harley, P.; Fall, R. (2000) Natural emissions of non-methane volatile
4 organic compounds, carbon monoxide, and oxides of nitrogen from North America. Atmos. Environ.
5 34:2205-2230.
6 Haagen-Smit, A. J. (1952) Chemistry and physiology of Los Angeles smog. Ind. Eng. Chem. 44: 1342-1346.
7 Henry, R. C. (1997) History and fundamentals of multivariate air quality receptor models. Chemom. Intell. Lab.
8 Syst. 37: 37-42.
9 Hewitt, C. N. (2001) The atmospheric chemistry of sulphur and nitrogen in power station plumes. Atmos. Environ.
10 35: 1155-1170.
11 Hidy, G. M. (1994) Atmospheric sulfur and nitrogen oxides: eastern North American source-receptor relationships.
12 San Diego, CA: Academic Press, Inc.
13 Hildemann, L. M.; Rogge, W. F.; Cass, G. R.; Mazurek, M. A.; Simoneit, B. R. T. (1996) Contribution of primary
14 aerosol emissions from vegetation-derived sources to fine particle concentrations in Los Angeles. J.
15 Geophys. Res. [Atmos.] 101: 19,541-19,549.
16 Hoffmann, T.; Odum, J. R.; Bowman, F.; Collins, D.; Klockow, D.; Flagan, R. C.; Seinfeld, J. H. (1997) Formation
17 of organic aerosols from the oxidation of biogenic hydrocarbons. J. Atmos. Chem. 26: 189-222.
18 Hopke, P. K.; Wlaschin, W.; Landsberger, S.; Sweet, C.; Vermette, S. J. (1988) The source apportionment of PM10
19 in south Chicago. In: Mathai, C. V.; Stonefield, D. H., eds. PM.10: implementation of standards, an
20 APCA/EPA international specialty conference; February; San Francisco, CA. Pittsburgh, PA: Air Pollution
21 Control Association; pp. 484-494. (APCA transactions series TR-13).
22 Houck, J. E.; Rau, J. A.; Body, S.; Chow, J. C. (1992) Source apportionment: Pocatello, Idaho PM10 nonattainment
23 area. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate source controls, an
24 A&WMA/EPA international specialty conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air &
25 Waste Management Association; pp. 219-230. (A&WMA transactions series no. 22).
26 Huang, X.; Olmez, I.; Aras, N. K.; Gordon, G. E. (1994) Emissions of trace elements from motor vehicles: potential
27 marker elements and source composition profile. Atmos. Environ. 28: 1385-1391.
28 Husar, R. B.; Schichtel, B. A.; Falke, S. R.; Li, F.; Wilson, W. E.; Pinto, J.; Malm, W. C.; Fox, D. G.; Feldman,
29 G. C.; McClain, C.; Kuring, N.; Holben, B. N.; Vermote, E. F.; Herman, J. R.; Elvidge, C. D. (2000) The
3 0 impact of the 1998 Central American smoke on the atmospheric environment of eastern North America.
31 J. Geophys. Res.: submitted.
32 Inhalable Paniculate Network. (1985) [Inhalable Paniculate Network Database.] Research Triangle Park, NC:
33 U.S. Environmental Protection Agency, National Exposure Research Laboratory; Jose M. Sune, contact.
34 Izumi, K.; Fukuyama, T. (1990) Photochemical aerosol formation from aromatic hydrocarbons in the presence of
35 NOX. Atmos. Environ. Part A 24: 1433-1441.
36 Jang, M.; Kamens, R. M. (1999) Newly characterized products and composition of secondary aerosols from the
37 reaction of a-pinene with ozone. Atmos. Environ. 33: 459-474.
38 Jang, M.; Kamens, R. M. (200la) Atmospheric secondary aerosol formation by heterogeneous reactions of
39 aldehydes in the presence of a sulfuric acid aerosol catalyst. Environ. Sci. Technol. 35: 4758-4766.
40 Jang, M.; Kamens, R. M. (200 Ib) Characterization of secondary aerosol from the photooxidation of toluene in the
41 presence of NOX and 1-propene. Environ. Sci. Technol. 35: 3626-3639.
42 Kamens, R.; Jang, M.; Chien, C.-J.; Leach, K. (1999) Aerosol formation from the reaction of a-pinene and ozone
43 using a gas-phase kinetics-aerosol partitioning model. Environ. Sci. Technol. 33: 1430-1438.
44 Kao, A. S.; Friedlander, S. K. (1995) Frequency distributions of PM10 chemical components and their sources.
45 Environ. Sci. Technol. 29: 19-28.
46 Kavouras, I. G.; Mihalopoulos, N.; Stephanou, E. G. (1998) Formation of atmospheric particles from organic acids
47 produced by forests. Nature (London) 395: 683-686.
48 Kavouras, I. G.; Mihalopoulus, N.; Stephanou, E. G. (1999a) Formation and gas/particle partitioning of
49 monoterpene photooxidation products over forests. Geophys. Res. Lett. 26: 55-58.
50 Kavouras, I. G.; Mihalopoulus, N.; Stephanou, E. G. (1999b) Secondary organic aerosol formation vs primary
51 organic aerosol emission: in situ evidence for the chemical coupling between monoterpene acidic
52 photooxidation products and new particle formation over forests. Environ. Sci. Technol. 33: 1028-1037.
53 Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
54 segregated mass concentration data and ultrafine particle number concentrations in urban areas.
55 Atmos. Environ. 33: 2907-2913.
April 2002 3-97 DRAFT-DO NOT QUOTE OR CITE
-------
1 Khalil, M. A. K.; Rasmussen, R. A. (1992) Forest hydrocarbon emissions: relationships between fluxes and ambient
2 concentrations. J. Air Waste Manage. Assoc. 42: 810-813.
3 Kim, B. M.; Zeldin, M. D.; Liu, C. S. (1992) Source apportionment study for state implementation plan
4 development in the Coachella Valley. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional
5 paniculate source controls: an A&WMA/EPA international specialty conference, v. II; January; Scottsdale,
6 AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 979-991. (A&WMA transactions series
7 no. 22).
8 Kinney, P. L.; Aggarwal, M.; Northridge, M. E.; Janssen, N. A. H.; Shepard, P. (2000) Airborne concentrations of
9 PM2.5 and diesel exhaust particles on Harlem sidewalks: a community-based pilot study. Environ. Health
10 Perspect. 108: 213-218.
11 Koch, S.; Winterhalter, R.; Uherek, E.; Kolloff, A.; Neeb, P.; Moortgat, G. K. (2000) Formation of new particles in
12 the gas-phase ozonolysis of monoterpenes. Atmos. Environ. 34: 4031-4042.
13 Koutrakis,P. (1995) [PM2 5 and PM10 data for Philadelphia, 1992-1995]. Boston, MA: Harvard School of Public
14 Health.
15 Landis, M. S.; Norris, G. A.; Williams, R. W.; Weinstein, J. P. (2001) Personal exposures to PM2 5 mass and trace
16 elements in Baltimore, MD, USA. Atmos. Environ. 35: 6511-6524.
17 Larsen, B. R.; Di Bella, D.; Glasius, M.; Winterhalter, R.; Jensen, N. R.; Hjorth, J. (2001) Gas-phase OH oxidation
18 of monoterpenes: gaseous and paniculate products. J. Atmos. Chem. 38: 231-276.
19 Leach, K. B.; Kamens, R. M.; Strommen, M. R.; Jang, M. (1999) Partitioning of semivolatile organic compounds in
20 the presence of a secondary organic aerosol in a controlled atmosphere. J. Atmos. Chem. 33: 241-264.
21 Leaderer, B. P.; Naeher, L.; Jankun, T.; Balenger, K.; Holford, T. R.; Toth, C.; Sullivan, J.; Wolf son, J. M.;
22 Koutrakis, P. (1999) Indoor, outdoor, and regional summer and winter concentrations of PM10, PM25, SO42",
23 H+, NH4+, NO3", NH3, and nitrous acid in homes with and without kerosene space heaters. Environ. Health
24 Perspect. 107:223-231.
25 Lin, J. J.; Noll, K. E.; Holsen, T. M. (1994) Dry deposition velocities as a function of particle size in the ambient
26 atmosphere. Aerosol Sci. Technol. 20: 239-252.
27 Lipfert, F. W. (1998) Trends in airborne paniculate matter in the United States. Appl. Occup. Environ. Hyg.
28 13:370-384.
29 Magliano, K. L.; Ranzieri, A.; Solomon, P.; Watson, J. (1998) Chemical mass balance modeling of data from the
30 1995 integrated monitoring study. Sacramento, CA: California Air Resources Board; California Regional
31 PM10/PM2.5 Air Quality Study.
32 Mazurek, M.; Masonjones, M. C.; Masonjones, H. D.; Salmon, L. G.; Cass, G. R.; Hallock, K. A.; Leach, M.
33 (1997) Visibility-reducing organic aerosols in the vicinity of Grand Canyon National Park: properties
34 observed by high resolution gas chromatography. J. Geophys. Res. [Atmos.] 102: 3779-3793.
35 McMurry, P. H.; Woo, K. S.; Weber, R.; Chen, D.-R.; Pui, D. Y. H. (2000) Size distributions of 3 to 10 nm
36 atmospheric particles: implications for nucleation mechanisms. Philos. Trans. R. Soc. London Ser. A
37 358:2625-2642.
3 8 Monn, C. (2001) Exposure assessment of air pollutants: a review on spatial heterogeneity and
39 indoor/outdoor/personal exposure to suspended paniculate matter, nitrogen dioxide and ozone. Atmos.
40 Environ. 35: 1-32.
41 Motallebi, N. (1999) Wintertime PM25 and PM10 source apportionment at Sacramento, California. J. Air Waste
42 Manage. Assoc. 49: PM-25-34.
43 NARSTO. (2002) Fine particle assessment (2nd review draft). Available: http://www.cgenv.com/Narsto/
44 (13 February 2002).
45 Neas, L. M. (1996) [Data provided for 1986 from Harvard six cities data base.] Boston, MA: Harvard School of
46 Public Health.
47 Norris, G. (1998) Air pollution and the exacerbation of asthma in an arid, western, US city [dissertation]. Seattle,
48 WA: University of Washington. Available from: University Microfilms International, Ann Arbor, MI;
49 publication no. 9836228.
50 Norris, G.; Young-Pong, S. N.; Koenig, J. Q.; Larson, T. V; Sheppard, L.; Stout, J. W. (1999) An association
51 between fine particles and asthma emergency department visits for children in Seattle. Environ. Health
52 Perspect. 107: 489-493.
53 Odum, J. R.; Hoffmann, T.; Bowman, F.; Collins, D.; Flagan, R. C.; Seinfeld, J. H. (1996) Gas/particle partitioning
54 and secondary organic aerosol yields. Environ. Sci. Technol. 30: 2580-2585.
April 2002 3-98 DRAFT-DO NOT QUOTE OR CITE
-------
1 Odum, J. R.; Jungkamp, T. P. W.; Griffin, R. I; Flagan, R. C.; Seinfeld, J. H. (1997a) The atmospheric
2 aerosol-forming potential of whole gasoline vapor. Science (Washington, DC) 276: 96-99.
3 Odum, J. R.; Jungkamp, T. P. W.; Griffin, R. J.; Forstner, H. J. L.; Flagan, R. C.; Seinfeld, J. H. (1997b) Aromatics,
4 reformulated gasoline, and atmospheric organic aerosol formation. Environ. Sci. Technol. 31: 1890-1897.
5 Ondov, J. M. (1996) Paniculate tracers for source attribution: potential for application to California's San Joaquin
6 Valley. J. Aerosol Sci. 27(suppl. 1): S687-S688.
7 Ott, S.-T.; Ott, A.; Martin, D. W.; Young, J. A. (1991) Analysis of a trans-Atlantic Saharan dust outbreak based on
8 satellite and GATE data. Mon. Weather Rev. 119: 1832-1850.
9 Paatero, P.; Tapper, U. (1993) Analysis of different modes of factor analysis as least squares fit problems. Chemom.
10 Intelligent Lab. Syst. 18: 183-194.
11 Paatero, P.; Tapper, U. (1994) Positive matrix factorization: a non-negative factor model with optimal utilization of
12 error estimates of data values. Environmetrics 5: 111-126.
13 Pandis, S. N.; Paulson, S. E.; Seinfeld, J. H.; Flagan, R. C. (1991) Aerosol formation in the photooxidation of
14 isoprene and p-pinene. Atmos. Environ. Part A 25: 997-1008.
15 Pandis, S. N.; Harley, R. A.; Cass, G. R.; Seinfeld, J. H. (1992) Secondary organic aerosol formation and transport.
16 Atmos. Environ. Part A 26: 2269-2282.
17 Perry, K. D.; Cahill, T. A.; Eldred, R. A.; Dutcher, D. D.; Gill, T. E. (1997) Long-range transport of North African
18 dust to the eastern United States. J. Geophys. Res. [Atmos] 102: 11,225-11,238.
19 Pinto, J. P.; Pleasant, M.; Kellogg, R.; Torok, Sz.; Stiles, D.; Willis, R. D.; Koutrakis, P.; Allen, G. A. (1995)
20 Chemical characterization of the ambient aerosol in Philadelphia. Presented at: Paniculate matter: health and
21 regulatory issues: an A&WMA international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA:
22 Air & Waste Management Association; submitted abstract.
23 Pinto, J. P.; Stevens, R. K.; Willis, R. D.; Kellogg, R.; Mamane, Y.; Novak, J.; Santroch, J.; Benes, L; Lenicek, J.;
24 Bures, V. (1998) Czech air quality monitoring and receptor modeling study. Environ. Sci. Technol. 32:
25 843-854.
26 Pinto, J. P.; Lefohn, A. S.; Shadwick, D. S. (2002) Aspects of the spatial vanability of PM25 concentrations within
27 urban areas of the United States. Environ. Sci. Technol.: submitted.
28 Pio, C.; Alves, C.; Duarte, A. (2001) Organic components of aerosols in a forested area of central Greece.
29 Atmos. Environ. 35: 389-401.
30 Placet, M.; Battye, R. E.; Fehsenfeld, F. C.; Bassett, G. W. (1991) Emissions involved in acidic deposition
31 processes. In: Irving, P. M., ed. Acidic deposition: state of science and technology, volume I: emissions,
32 atmospheric processes, and deposition. Washington, DC: The U.S. National Acid Precipitation Assessment
33 Program. (State of science and technology report no. 1).
34 Poore, M. W. (2000) Oxalic acid in PM2 5 paniculate matter in California. J. Air Waste Manage. Assoc.
35 50: 1874-1875.
36 Prospero, J. M. (1996) The atmospheric transport of particles to the ocean. In: Ittekkot, V.; Honjo, S.; Depetris,
37 P. J., eds. Particle flux in the ocean. New York, NY: John Wiley & Sons, Ltd.; pp. 19-52. (SCOPE report
38 no. 57).
39 Prospero, J. M. (1999) Long-range transport of mineral dust in the global atmosphere: impact of African dust on the
40 environment of the southeastern United States. Proc. Natl. Acad. Sci. U.S.A. 96: 3396-3403.
41 Prospero, J. M.; Nees, R. T.; Uematsu, M. (1987) Deposition rate of paniculate and dissolved aluminum derived
42 from Saharan dust in precipitation at Miami, Florida. J. Geophys. Res. [Atmos.] 92: 14,723-14,731.
43 Prospero, J. M.; Olmez, I.; Ames, M. (2001) Al and Fe in PM2.5 and PM10 suspended particles in south-central
44 Florida: the impact of the long range transport of African mineral dust. J. Water Air Soil Pollut.
45 125:291-317.
46 Pun, B. K.; Seigneur, C.; Grosjean, D.; Saxena, P. (2000) Gas-phase formation of water-soluble organic compounds
47 in the atmosphere: a retrosynthetic analysis. J. Atmos. Chem. 35: 199-223.
48 Ramadan, Z.; Song, X.-H.; Hopke, P. K. (2000) Identification of sources of Phoenix aerosol by positive matrix
49 factorization. J. Air Waste Manage. Assoc. 50: 1308-1320.
50 Rizzo, M.; Pinto, J. P. (2001) Initial characterization of fine paniculate matter (PM2.5) collected by the National
51 Federal Reference Monitoring Network. Presented at: 94th annual conference & exhibition of the Air
52 &Waste Management Association; June, Orlando, FL. Pittsburgh, PA: Air & Waste Management
53 Association.
54 Rodes, C. E.; Evans, E. G. (1985) Preliminary assessment of 10 /j.m paniculate sampling at eight locations in the
55 United States. Atmos. Environ. 19: 293-303.
April 2002 3-99 DRAFT-DO NOT QUOTE OR CITE
-------
1 Rogge, W. F.; Hildemann, L. M; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993) Sources of fine organic
2 aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci.
3 Technol. 27:636-651.
4 Ryan, W. M.; Badgett-West, C. R.; Holtz, D. R.; Peters, T. A.; Cooper, J. A.; Ono, D. (1988) Reconciliation of
5 receptor and dispersion modeling impacts of PM10 in Hayden, Arizona. In: Mathai, C. V.; Stonefield, D. H.,
6 eds. PM.10: implementation of standards, an APCA/EPA international specialty conference; February;
7 San Francisco, CA. Pittsburgh, PA: Air Pollution Control Association; pp. 419-429. (APCA transactions
8 series TR-13).
9 Salma, L; Maenhaut, W.; Zaray, G. (2002) Comparative study of elemental mass size distributions in urban
10 atmospheric aerosol. J. Aerosol Sci. 33: 339-356.
11 Savoie, D. L.; Prospero, J. M. (1980) Water-soluble potassium, calcium, and magnesium in the aerosols over the
12 tropical North Atlantic. J. Geophys. Res. 85: 385-392.
13 Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
14 literature and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem. 24: 57-109.
15 Schauer, J. J.; Cass, G. R. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants
16 using organic compounds as tracers. Environ. Sci. Technol. 34: 1821-1832.
17 Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurik, M. A.; Cass, G. R. (1996) Source apportionment of
18 airborne paniculate matter using organic compounds as tracers. Atmos. Environ. 30: 3837-3855.
19 Shinn, E. A.; Smith, G. W.; Prospero, J. M.; Betzer, P.; Hayes, M. L.; Garrison, V.; Barber, R. T. (2000) African
20 dust and the demise of Caribbean coral reefs. Geophys. Res. Lett. 27: 3029-3032.
21 Skidmore, L. W.; Chow, J. C.; Tucker, T. T. (1992) PM10 air quality assessment forthe Jefferson County, Ohio air
22 quality control region. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional paniculate source
23 controls, an A&WMA/EPA international specialty conference, v. II; January; Scottsdale, AZ. Pittsburgh, PA:
24 Air and Waste Management Association; pp. 1016-1031. (A&WMA transactions series no. 22).
25 Smith, G. T.; Ives, L. D.; Nagelkerken, I. A.; Ritchie, K. B. (1996) Caribbean sea fan mortalities. Nature (London)
26 383:487.
27 Spengler, J. D.; Briggs, S. L. K.; Ozkaynak, H. (1986) Relationship between TSP measurements and
28 size-fractionated particle mass measurements in six cities participating in the Harvard Air Pollution Health
29 Study. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning
30 and Standards; EPA cooperative agreement no. CR-812667-01-0; December 5.
31 Spengler, J. D.; Koutrakis, P.; Dockery, D. W.; Raizenne, M.; Speizer, F. E. (1996) Health effects of acid aerosols
32 on North American children: air pollution exposures. Environ. Health Perspect. 104: 492-499.
33 Suarez, A. E.; Caffrey, P. F.; Ondov, J. M.; Thaung, K. C. (1996) Size distribution of Ir-tagged soot from diesel
34 sanitation trucks. J. Aerosol Sci. 27(suppl. 1): S697-S698.
35 Tanner, R. L.; Parkhurst, W. J. (2000) Chemical composition of fine particles in the Tennessee Valley region. J. Air
36 Waste Manage. Assoc. 50: 1299-1307.
37 Taylor, C. A., Jr.; Stover, C. A.; Westerdahl, F. D. (1998) Speciated fine particle (<2.5 ^m aerodynamic diameter)
38 and vapor-phase acid concentrations in southern California. Presented at: 91st annual meeting & exhibition
39 of the Air & Waste Management Association; June; San Diego, CA. Pittsburgh, PA: Air & Waste
40 Management Association; paper no. WA7401.
41 Thanukos, L. C.; Miller, T.; Mathai, C. V.; Reinholt, D.; Bennett, J. (1992) Intercomparison of PM10 samplers and
42 source apportionment of ambient PM10 concentrations at Rillito, Arizona. In: Chow, J. C.; Ono, D. M., eds.
43 PM10 standards and nontraditional paniculate source controls, an A&WMA/EPA international specialty
44 conference, v. I; January; Scottsdale, AZ. Pittsburgh, PA: Air & Waste Management Association;
45 pp. 244-261. (A&WMA transactions series no. 22).
46 Tobias, H. J.; Ziemann, P. J. (2000) Thermal desorption mass spectrometric analysis of rganic aerosol formed from
47 reactions of 1-tetradecene and O3 in the presence of alcohols and carboxylic acids. Environ. Sci. Technol.
48 34:2105-2115.
49 Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin:
50 a descriptive analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
51 Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
52 primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
53 Turpin, B. J.; Lim, H.-J. (2001) Species contributions to PM2.5 mass concentrations: revisiting common
54 assumptions for estimating organic mass. Aerosol Sci. Technol. 35: 602-610.
April 2002 3-100 DRAFT-DO NOT QUOTE OR CITE
-------
1 Turpin, B. J.; Huntzicker, J. J.; Larson, S. M.; Cass, G. R. (1991) Los Angeles summer midday paniculate carbon:
2 primary and secondary aerosol. Environ. Sci. Technol. 25: 1788-1793.
3 U.S. Environmental Protection Agency. (1995) Compilation of air pollutant emission factors. Volume I: stationary
4 point and area sources. 5th ed. Research Triangle Park, NC: Office of Air Quality Planning and Standards;
5 report no. AP-42-ED-5. Available from: NTIS, Springfield, VA; PB95-196028.
6 U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
7 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
8 U.S. Environmental Protection Agency. (1999) Getting started: emissions inventory methods for PM-2.5. Research
9 Triangle Park, NC: Office of Air Quality Planning and Standards, Emission Factor and Inventory Group;
10 EIIP document series - volume IX. Available:
11 http://www.epa.gov/ttn/chief/eiip/techreport/volume09/index.html [4 April, 2002].
12 U.S. Environmental Protection Agency. (2000a) Aerometric Information Retrieval System (AIRS). Research
13 Triangle Park, NC: Office of Air Quality Planning and Standards. Available: www.epa.gov/ttn/airs/
14 [2000, October 6].
15 U.S. Environmental Protection Agency. (2000b) Air quality criteria for carbon monoxide. Research Triangle Park,
16 NC: National Center for Environmental Assessment; report no. EPA/600/P-99/001F. Available:
17 www.epa.gov/ncea/co/ [2000, October 6].
18 U.S. Environmental Protection Agency. (2000c) National air pollutant emission trends, 1900-1998. Research
19 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA-454/R-00-002. Available:
20 http://www.epa.gov/ttn/chief/trends/trends98/ [2001, January 30].
21 U.S. Environmental Protection Agency. (2000d) National air quality and emissions trends report, 1998. Research
22 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/454/R-00-003. Available
23 from: NTIS, Springfield, VA; PB2000-104960.
24 U.S. Environmental Protection Agency. (2001) National air quality and emissions trends report, 1999. Research
25 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/454/R-01-004. Available:
26 http://www.epa.gov/oar/aqtrnd99/ [11 March, 2002].
27 Vermette, S. J.; Williams, A. L.; Landsberger, S. (1992) PM10 source apportionment using local surface dust
28 profiles: examples from Chicago. In: Chow, J. C.; Ono, D. M., eds. PM10 standards and nontraditional
29 paniculate source controls, an A&WMA/EPA international specialty conference, v. I; January; Scottsdale,
30 AZ. Pittsburgh, PA: Air & Waste Management Association; pp. 262-278. (A&WMA transactions series
31 no. 22).
32 Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of
33 submicronparticle exposure. Environ. Health Perspect. 108: 1139-1145.
34 Watson, J. G.; Chow, J. C. (1999) Reconciling urban fugitive dust emissions inventory and ambient source
3 5 contribution estimates: summary of current knowledge and needed research [draft]. Reno, NV: Desert
36 Research Institute; DRI document no. 6110.4D2.
37 Watson, J. G.; Chow, J. C. (2001) Estimating middle-, neighborhood-, and urban-scale contributions to elemental
3 8 carbon in Mexico City with a rapid response aethalometer. J. Air Waste Manage. Assoc. 51: 1522-1528.
39 Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
40 (1990a) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
41 Watson, J. G.; Chow, J. C.; Pritchett, L. C.; Houck, J. A.; Ragazzi, R. A.; Burns, S. (1990b) Chemical source
42 profiles for paniculate motor vehicle exhaust under cold and high altitude operating conditions. Sci. Total
43 Environ. 93: 183-190.
44 Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M.; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994)
45 Chemical mass balance source apportionment of PM10 during the Southern California Air Quality Study.
46 Aerosol Sci. Technol. 21: 1-36.
47 West, J. J.; Ansari, A. S.; Pandis, S. N. (1999) Marginal PM2 5: nonlinear aerosol mass response to sulfate
48 reductions in the eastern United States. J. Air Waste Manage. Assoc. 49: 1415-1424.
49 White, W. H. (1998) Statistical considerations in the interpretation of size-resolved paniculate mass data. J. Air
50 Waste Manage. Assoc. 48: 454-458.
51 Wilson, W. E.; Stockburger, L. (1990) Diurnal variations in aerosol composition and concentration. In: Masuda, S.;
52 Takahashi, K., eds. Aerosols: science, industry, health and environment: proceedings of the third
53 international aerosol conference, v. 2; September; Kyoto, Japan. Oxford, United Kingdom: Pergamon Press
54 pic; pp. 962-965.
April 2002 3-101 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
2 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
3 Winer, A. M; Arey, J.; Atkinson, R.; Aschmann, S. M; Long, W. D.; Morrison, C. L.; Olszyk, D. M. (1992)
4 Emission rates of organics from vegetation in California's Central Valley. Atmos. Environ. Part A
5 26:2647-2659.
6 Winterhalter, R.; Neeb, P.; Grossmann, D.; Kolloff, A.; Horie, O.; Moortgat, G. (2000) Products and mechanisms of
7 the gas phase reaction of ozone with p-pinene. J. Atmos. Chem. 35: 165-197.
8 Wongphatarakul, V.; Friedlander, S. K.; Pinto, J. P. (1998) A comparative study of PM25 ambient aerosol chemical
9 databases. Environ. Sci. Technol. 32: 3926-3934.
10 Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; McMurry, P. H. (2001) Measurement of Atlanta aerosol size distributions:
11 observations of ultrafine particle events. Aerosol Sci. Technol. 34: 75-87.
12 Wotawa, G.; Trainer, M. (2000) The influence of Canadian forest fires on pollutant concentrations in the United
13 States. Science (Washington, DC) 288: 324-328.
14 Yu, J.; Cocker, D. R., Ill; Griffin, R. J.; Flagan, R. C.; Seinfeld, J. H. (1998) Gas-phase ozone oxidation of
15 monoterpenes: gaseous and paniculate products. J. Atmos. Chem. 34: 207-258.
16 Yu, J.; Griffin, R. J.; Cocker, D. R., Ill; Flagan, R. C.; Seinfeld, J. H. (1999) Observation of gaseous and paniculate
17 products of monoterpene oxidation in forest atmospheres. Geophys. Res. Lett. 26: 1145-1148.
18 Zhang, S.-H.; Shaw, M.; Seinfeld, J. H.; Flagan, R. C. (1992) Photochemical aerosol formation from a-pinene and
19 P-pinene. J. Geophys. Res. [Atmos.] 97: 20,717-20,729.
20 Zweidinger, R.; Purdue, L.; Fitzgerald, K.; Carmichael, L.; Kellog, R. (1998) NERL PM research monitoring
21 platforms: Baltimore, Fresno and Phoenix, data report. Research Triangle Park, NC: U.S. Environmental
22 Protection Agency, National Exposure Research Laboratory; report no. EPA/600/R-98/138. Available from:
23 NTIS, Springfield, VA; PB99-105660.
April 2002 3-102 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 3A
2
3 Spatial and Temporal Variability of the Nationwide AIRS
4 PM2 5 and PM10.2 5 Data Sets
7 Aspects of the spatial and temporal variability of 24-h average PM25 concentrations for
8 1999 and 2000 in a number of metropolitan statistical areas (MS As) across the United States are
9 presented in this Appendix. Data for multiple sites in 27 urban areas have been obtained from
10 the AIRS data base and analyzed for their seasonal variations, for their spatial correlations, and
11 for their spatial uniformity (Pinto et al., 2002). A number of aspects of the spatial and temporal
12 variability of the PM2 5 data set from 1999 were presented in Rizzo and Pinto (2001), based in
13 part on analyses given in Fitz-Simons et al. (2000).
14 Quality assured measurements for at least fifteen days during each calendar quarter for
15 1999 and 2000, or for 2000 alone, at a minimum of four monitoring sites in a given MSA were
16 required for their inclusion in the analyses given in this appendix. The Baton Rouge, LA MSA,
17 which had only three sites meeting this criterion, was an exception. Data from Baton Rouge
18 were included for the sake of geographic coverage. Typically, at least 200 measurements were
19 available for each monitoring site chosen. Monitoring sites were chosen without consideration of
20 the land use type used to characterize their locations.
21 Because of changes in monitoring strategies, funding levels etc., there were year to year
22 changes in monitoring sites meeting the above criteria in a number of MS As. Data for the
23 Philadelphia, PA, Norfolk, VA, Pittsburgh, PA, Detroit, MI, Chicago, IL, Louisville, KY,
24 St. Louis, MO, and the Dallas, TX MSAs have been analyzed only for the year 2000 because of a
25 lack of consistent coverage in 1999.
26 Information about seasonal and spatial variability in PM2 5 concentrations within 27 MSAs
27 across the United States are provided in the accompanying figures (Figures 3A-1 to 3A-27).
28 Underneath the value for r, the 90th percentile values of the absolute difference in PM2 5
29 concentrations (in //g/m3) and the coefficient of divergence (COD) are given in parentheses.
April 2002 3 A-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 Underneath these two measures of spatial variability, the numbers of observations used in the
2 calculations of the statistics in part c of each figure are given.
3 The COD was defined mathematically and used earlier in Chapter 3 as a measure of the
4 degree of similarity between two data sets. A COD of zero implies that values in both data sets
5 are identical, and a COD of one indicates that two data sets are completely different. Values of
6 P90 provide a measure in absolute terms of differences in concentrations between sites, and CODs
7 provide a relative measure of these differences. The maximum number of days of coincident
8 data from paired sites were used to calculate correlation coefficients, values for P90, and CODs.
9 The correlation coefficients were also calculated by using only concurrent measurements
10 obtained at all of the monitoring sites within urban areas meeting the above selection criteria.
11 The correlation coefficients that were calculated differed only in the third significant figure
12 between the two methods.
13 Information about the spatial and temporal variability of 24-h average PM10_25
14 concentrations is summarized in Figures 3A-28 to 3A-32. Data are shown for 2000 for all
15 MS As, except the Los Angeles-Long Beach MSA, for which data are shown for 1999.
16 A schematic map showing locations of sampling sites within each MSA is given in part a, at the
17 top of each figure. Also included in the map are major highways and a distance scale. A key
18 giving the AIRS site ID #'s is shown alongside each map. Box plots showing lowest, lower
19 quartile, median, upper quartile and highest PM2 5 concentrations for each calendar quarter are
20 shown in part b of each figure. AIRS site ID #'s, annual mean concentrations, the number of
21 observations, and the standard deviation of the data are shown above the box plots. Finally, in
22 part c of each figure, statistics characterizing the spatial variability in PM2 5 concentrations are
23 given. For each site-pair, the Pearson correlation coefficient (r) is provided. Underneath each
24 value for r, the number of observations is given.
25
April 2002 3 A-2 DRAFT-DO NOT QUOTE OR CITE
-------
Philadelphia, PA MSA
a.
Site A
SiteB
SiteC
SiteD
SiteE
AIRS Site ID
34-007-0003
34-007-1007
42-045-0002
42-101-0004
42-101-0136
50
100 km
J. AIRS ID#
Mean
Obs
SD
3400700031 3400710071 4204500021 4210100041 4210101361
50-
40-
30-
20-
10-
0-
c. Site
14.941
108
8.524
15.427
103
8.749
15.992
112
8.265
14.823
284
8.537
14.718
278
8.295
1234
B
A
B
C
D
E
0.964 0.
1 (3.3, 0.082) (6.3,
95
,868
0.155)
98
0.849
1 (6.9,0.158)
94
1
0.88
(4.8, 0.129)
81
0.894
(3.7, 0.135)
77
0.868
(5.0, 0.149)
85
1
0.868
(5.4, 0.147)
80
0.857
(6.4, 0.148)
79
0.818
(6.6,0.154)
83
0.918
(4.9,0.13)
246
1
Figure 3A-1. Philadelphia, PA-NJ MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM2S concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-3
DRAFT-DO NOT QUOTE OR CITE
-------
Washington, DC MSA
a.
b.
( J^S ;;^W!
7'X/ ~~^rQ'~'\ Site A
/ y B~7~-~\^-j£~'~- ^'*e ^
C"I:"'~~~7- **/ X^Eote NV Site C
/^.^W'^^^^^i^JSpZ'^ /' Site D
>>. •• '(^w>fr\7 'p~i Site E
"\ \_f /Y /^VAs
'c.. / ^"''J 0 50 100km
AIRSIDO 1100100411 1100100431 5101300201 5105900301 5110710051
Mean
Obs
SD
125 -
100 -
75-
50-
25-
0-
16.298 15.075 14.670 12.891 13.602
520 575 211 549 209
9.634 8.291 7.921 7.344 8.248
I I • I
1 11 ll 1
III 1 1 III! II 1 Illl
1234 1234 1234 1234 1234
c. Site ABODE
0.77 0.872 0.756 0.833
A 1 (6.1,0.216) (5.8,0.168) (7.7,0.196) (7.4,0.207)
453 157 410 159
0.875 0.867 0.849
B 1 (6.40.210) (6.6,0.197) (7.6,0.238)
175 449 174
0.927 0.938
C 1 (4.2,0.096) (4.9,0.115)
175 198
0.868
D 1 (5.9,0.124)
171
E 1
F
AIRS Site ID
11-001-0041
11-001-0043
51-013-0020
51-059-0030
51-107-1005
54-003-0003
5400300031
16.278
203
9.258
III
1234
F
0.736
(10.1,0.215)
158
0.787
(7.8, 0.227)
175
0.841
(8.2, 0.161)
182
0.791
(9.6, 0.189)
156
0.876
(8.6, 0.168)
179
1
Figure 3A-2. Washington, DC MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-4
DRAFT-DO NOT QUOTE OR CITE
-------
a.
Norfolk, VA MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
51-550-0012
51-650-0004
51-700-0013
51-710-0024
51-810-0008
30
60 km
AIRSID* 5155000121 5165000041 5170000131 5171000241 5181000081
c.
Mean
Obs
SD
50-
40-
30-
20-
10-
0-
Site
13.801
284
6.987
13.408
104
7.034
13.134
113
6.761
13.835
117
7.001
13.180
116
6.997
1234 123
B
D
A
B
C
D
E
0.942 0.936
1 (4.2, 0.091) (4.7, 0.098)
77 85
0.974
1 (2.6, 0.068)
99
1
0.964
(3.5, 0.082)
88
0.969
(2.8, 0.067)
102
0.967
(3.4, 0.085)
108
1
0.953
(3.6, 0.077)
86
0.943
(3.5, 0.092)
100
0.935
(4.3, 0.091)
107
0.941
(3.6, 0.097)
113
1
Figure 3A-3. Norfolk, VA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-5
DRAFT-DO NOT QUOTE OR CITE
-------
Columbia, SC MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
45-063-0005
45-063-0008
45-079-0007
45-079-0019
40 km
1). AIRS IDS 4506300051 4506300081 4507900071 4507900191
Mean
Obs
SD
60
50
40
30
20
10
14.680
231
6.760
16.462
228
7.121
15.461
216
6.900
16.098
229
7.148
c.
Site
B
D
A
B
C
D
0.882
1 (5.3,0.121)
215
1
0.949
(3.9, 0.081)
204
0.933
(4.0, 0.082)
202
1
0.93
(4.8, 0.099)
216
0.949
(3.3, 0.067)
216
0.971
(2.7, 0.06)
203
1
Figure 3A-4. Columbia, SC MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM2 5 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-6
DRAFT-DO NOT QUOTE OR CITE
-------
Atlanta, GA MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
1 3-063-0091
1 3-089-0002
1 3-089-2001
13-121-0032
13-121-0039
13-121-1001
1 3-223-0003
40 80 km
C.
AIRSIDff
Mean
Obs
SD
150 -
125 -
100 -
75 -
50 -
25 -
0 -
Site
1306300911 1X8900021 1X8920011 1312100321
20.135 18.887 20.329 19.417
195 555 600 563
9.662 10.031 9.796 10.006
nil II I III! Illl
A B C D
0.808 0.828 0.795
A 1 (8.3,0.161) (8.1,0.115) (7.1,0.150)
148 167 158
0.77 0.669
B 1 (8.2 0.158) (8.3, 0.178)
466 440
0.727
C 1 (6.5,0.143)
474
D 1
E
F
G
1312100391 1312110011 1322300031
22.409 1 9.497
1 94 1 93
11.141 10.095
I I
1 7.869
198
1 3.557
III III! ill!
E F
G
0.82 0.886 0.543
(8.2,0.131) (7.7,0.140) (13.9,0.219)
159 162 158
0.855 0.841 0.639
(9.9,0.174) (9.3,0.181) (12.5,0.224)
150 148 155
0.861 0.802 0.574
(9.0, 0.129) (10.6, 0.192) (12.0, 0.207)
167 165 166
0.815 0.819 0.527
(9.4,0.157) (9.5,0.182) (10.90.218)
160 162 161
0.785 0.574
1 (10.5, 0.188) (14.0, 0.242)
158 160
1
0.547
(9.9, 0.213)
155
1
Figure 3A-5. Atlanta, GA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-7
DRAFT-DO NOT QUOTE OR CITE
-------
Birmingham, AL MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
01-730-0023
01-073-1005
01-732-2003
01-073-2006
01-073-5002
b.
AIRS IDS
Mean
Obs
SD
C.
80-
60
40-
20-
Site
60 km
0107300231 0107310051 0107320031 0107320061 0107350021
22.854
704
11.508
17.640
238
8.469
21.076
705
10.538
18.674
226
8.769
17.769
236
8.560
1234 1234 1234 1234 1234
B
D
A
B
C
D
E
0.807 0.86
1 (14.4, 0.205) (10.2, 0.154)
235 680
0.808
1 (9.3,0.180)
233
1
0.781
(14.2, 0.195)
223
0.858
(8.3, 0.160)
216
0.792
(10.6, 0.176)
222
1
0.8
(15.4, 0.196)
234
0.866
(7.5, 0.151)
228
0.811
(10.0, 0.167)
231
0.859
(8.1, 0.143)
214
1
Figure 3A-6. Birmingham, AL MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-8
DRAFT-DO NOT QUOTE OR CITE
-------
Tampa, FL MSA
a.
AIRS Site ID
AIRSID#
Mean
Obs
SD
Site A
Site B
SiteC
SiteD
12-057-0030
12-057-1075
12-103-0018
12-103-1008
C.
100 -I
80 -
60 -
40 -
20 -
0 H
30
60 km
1205700301 1205710751 1210300181 1210310081
13.200
651
5.872
I
12.710
654
6.326
':!
11
"
I'
|
12.160 11.599
682 222
6.292 6.501
I
ll
"
• II
i
I
ite
A
B
C
D
1234 1234
A B
0.733
1 (3.8,0.107)
593
1
1234
C
0.846
(4.5, 0.119)
604
0.631
(4.9, 0.142)
611
1
1234
D
0.852
(4.4, 0.118)
202
0.653
(5.0, 0.138)
203
0.786
(3.6, 0.103)
214
1
Figure 3A-7. Tampa, FL MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations; (c) Intersite correlation
statistics, for each data pair, the correlation coefficient, (P90, coefficient of
divergence) and number of measurements are given.
April 2002
3A-9
DRAFT-DO NOT QUOTE OR CITE
-------
Cleveland, OH MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
39-035-0013
39-035-0038
39-035-0060
39-035-0065
39-035-0066
39-035-1002
39-085-1001
40
80 km
I). AIRS ID# 3903500131 3903500381 3903500601 3903500651 3903500661 3903510021 3902510011
C.
Mean
Obs
SD
60
40-
20-
o-
18.879
226
9.669
20.273
18.823
228
9.9503
17.983
224
8.7837
14.806
216
8.2146
15.145
232
7.9503
13.851
222
7.7588
I
Site A
B
D
A
B
C
D
E
F
0.93 0.946
1 (6.9,0.110) (3.8,0.135)
194 206
0.904
1 (7.1,0.164)
192
1
0.939
(5.3, 0.096)
206
0.92
(6.8, 0.150)
192
0.869 0.793
(10.3, 0.156) (13. 8, 0.227)
189 184
0.912
(5.2, 0.156)
198
1
0.868
(8.3, 0.204)
189
0.94
(5.5, 0.161)
185
1
0.847
(9.7, 0.192)
198
0.791
(14.1, 0.220)
195
0.823
(9.6, 0.222)
197
0.868
(6.3, 0.179)
197
0.812
(5.7, 0.193)
192
1
0.872
(10.7,0.211)
197
0.793
(14.9, 0.240)
184
0.849
(11.8, 0.240)
192
0.874
(8.2, 0.216)
192
0.883
(6.6, 0.166)
185
0.85
(6.9, 0.196)
187
Figure 3A-8. Cleveland, OH MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-10
DRAFT-DO NOT QUOTE OR CITE
-------
Pittsburgh, PA MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
42-003-0008
42-003-0064
42-125-0200
42-125-5001
40
80 km
b.
c.
AIRS \D#
Mean
Obs
3D
80-
60-
40-
20-
0-
4200300081 4200300641 4212502001 4212550011
15.958 20.135 15.223 13.306
229 315 108 345
7.239 12.755 7.292 6.641
I
M
li
It I
I
i
1234 1234 1234 1234
Site A B C D
0.767 0.937 0.876
A 1 (14.2, 0.162) (4.1, 0.089) (7.40.154)
217 70 213
0.58 0.575
B 1 (21.3, 0.238) (20.5, 0.272)
95 297
0.905
C 1 (5.9,0.138)
98
D 1
Figure 3A-9. Pittsburgh, PA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3 A-11
DRAFT-DO NOT QUOTE OR CITE
-------
Steubenville, OH-Weirton, WV MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
39-081-0016
39-081-1001
54-009-0005
54-029-0011
54-029-0011
15
30 km
b.
AIRS IDS
Mean
Obs
SD
3908100161 3908110011 5400900051 5402900111 5402900112
60-
40-
19.262
214
10.031
18.256
523
8.465
17.261
222
7.717
16.522
229
10.082
16.760
220
10.333
1234
1234
c. Site A
B
D
A
B
C
D
E
0.831 0.85
1 (9.9,0.186) (8.7,0.172)
166 196
0.846
1 (6.6,0.171)
166
1
0.843
(8.2, 0.187)
201
0.86
(9.2,0.183)
174
0.883
(7.6, 0.164)
220
1
0.805
(9.9, 0.199)
194
0.816
(9.2, 0.207)
161
0.873
(7.9, 0.179)
204
0.978
(2.5, 0.101)
210
1
Figure 3A-10. Steubenville, OH-Weirton, WV MSA. (a) Locations of sampling sites by
AIRS ID#; (b) Quarterly distribution of 24-h average PM2 5 concentrations;
(c) Intersite correlation statistics, for each data pair, the correlation
coefficient, (P90, coefficient of divergence) and number of measurements are
given.
April 2002
3A-12
DRAFT-DO NOT QUOTE OR CITE
-------
Detroit, Ml MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
26-099-0009
26-125-0001
26-147-0005
26-163-0033
26-163-0036
40 km
C.
AIRS IDS
Mean
Obs
SD
50 •
40 •
30 •
20 •
10 •
o •
2609900091 2612500011 2614700051 2616300331 2616300361
13.450
113
7.922
I
I
I
I
I
1234
Site A
A 1
B
C
D
15.552 14.172
90 102
9.223 8.771
I
I
I
I
,
I
,
I
I
I
20.173 17.446
1 08 1 03
10.475 9.626
I
I
I
ll
|
|
I
|
I
I
1234 1234 1234 1234
B C
D E
0.958 0.952 0.931 0.926
(4.9,0.107) (5.6,0.127) (12.7, 0.222) (9.0, 0.177)
83 96 98 96
0.939 0.92 0.917
1 (5.8,0.121) (12.3 0.193) (8.3, 0.151)
73 77 75
1
0.876 0.875
(13.3, 0.222) (8.9, 0.197)
89 88
0.923
1 (7.1,0.108)
Figure 3A-11. Detroit MI MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM2 5 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-13
DRAFT-DO NOT QUOTE OR CITE
-------
Grand Rapids, Ml MSA
a.
k
AIRS Site ID
Site A
Site B
SiteC
Site D
26-005-0003
26-081-0020
26-121-0040
26-139-0005
30
60 km
b. AIRSID* 2600500031 2608100201 2612100401 2613900051
Mean
Obs
SD
50-
40-
30-
20-
10-
0-
11.939
674
8.078
13.821
696
8.422
11.922
257
8.295
13.083
235
8.543
c.
Site
B
A
B
C
D
0.935
1 (5.5, 0.152)
647
1
0.924
(4.6, 0.135)
237
0.932
(6.1, 0.162)
244
1
0.954
(3.9, 0.125)
222
0.984
(2.8, 0.086)
222
0.947
(5.0, 0.120)
208
1
Figure 3A-12. Grand Rapids, MI MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-14
DRAFT-DO NOT QUOTE OR CITE
-------
Milwaukee, Wl MSA
{, AIRS Site ID
f~f Site A 55-079-0010
a.
f f SiteB 55-079-0026
V SiteC 55-079-0043
n Site D fis-n?q-nnsi
V Site E 55-079-0059
, ^_ K) SiteF 55-079-0099
| -»— ~^p^_ f9*ip SiteG 55-133-0027
JJ-®*i\ SiteH 55-133-0034
>-l«' TO 1.
S J \ 1 1
0 20 40 km
AIRS ID# 5507900102 5507900261 5507900431 5507900511 5507900592 5507900991 5513300272 5513300341
Mean
Obs
SD
60-
50-
40-
30-
20-
10-
0-
13.930 13.181 14.688 13.082 14.315 14.222 13.723 13.088
698 685 220 216 224 235
472 234
8.409 8.302 8.990 8.250 8.154 8.959 7.861 7.903
::::::
1. ,1 ; 1,. 1
III HI HI II HI 1
1234 1234 1234 1234 1234 1234 1
c. Site A B C D E F
1
[
,11 1
234 1234
G H
0.934 0.966 0.974 0.96 0.977 0.96 0.961
A 1 (3.3,0.111) (3.7,0.089) (3.0,0.086) (3.0,0.094) (3.9,0.085) (3.5
0.089) (3.3,0.108)
654 213 209 217 228 451 227
0.947 0.947 0.944 0.963 0.95 0.932
B 1 (3.7,0.112) (3.8,0.116) (3.8,0.112) (3.8,0.110) (3.7
0.111) (3.6,0.130)
210 206 215 225 441 224
0.945 0.95 0.964 0.948 0.929
C 1 (4.4,0.121) (4.7,0.100) (3.60.094) (3.9
0.106) (5.3 0.141)
188 187 198 212 200
0.954 0.956 0.967 0.964
D 1 (3.9,0.111) (4.5,0.114) (3.6
0.093) (3.5,0.111)
194 212 207 209
0.963 0.959 0.96
E 1 (3.5, 0.098) (3.5
0.095) (3.7 0.123)
203 211 207
0.963 0.95
F 1 (4.4
0.099) (5.5 0.133)
219 221
G
H
0.973
1 (2.9 0.101)
220
1
Figure 3A-13. Milwaukee, WI MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3 A-15
DRAFT-DO NOT QUOTE OR CITE
-------
Chicago, IL MSA
/ AIRS Site ID
a. ( Site A 17-031-0014
k^ ... |k SitfiR i7-rni-nn99
^-W— -Jft \ SiteC 17-031-0050
T~^L SiteD 17-031-0052
A Cfcr-*L SiteE 17-031-1016
-St—^r*1& A Qlto F <^-? n-a-i onn-i
r^n ® ^ SiteG 17-031-3301
Uf* ~~| SiteH 17-031-4006
IY Site I 17-031-4201
j « .. i
one j i /'-U43-4UU^
/ ' „'„ ' SiteK 17-197-1002
0.
AIRS ID# 1703100141 1703100221 1703100501 1703100521 170311 01 51 1703120011 1703133011 1703140061
Mean 1 5 823 1 7 933 1 6 996 1 8 295 20 277 1 6 790 1 6 889 1 5 268
Obs 104 113 274 346 108 113 115 101
SD 7.935 8.175 8.468 9.289 9.331 7.694 7.689 8.423
SC •'•':': '• '• '•
40 • :
I.
li
^ I ll '
II ll 1 1 i II 1
1234 1234 1234 1234 1234 1234 1234 1234
c. Site A B C D E F G H
0.912 0.946 0.909 0.921 0.902 0.927 0.876
A. 1 (4.4,0.121) (4.6,0.077) (6.6,0.13) (7.5,0.143) (5.6,0.111) (5.1,0.104) (5.8,0.133)
96 78 100 92 98 98 88
0.92 0.872 0.866 0.892 0.879 0.689
R 1 (5.4,0.113) (6.5,0.14) (7.0,0.141) (5.7,0.131) (6.0,0.132) (7.9,0.213)
87 108 103 104 106 92
0.941 0.93 0.955 0.923 0.75
C 1 (5.0,0.094) (7.8,0.012) (3.5,0.082) (5.3,0.096) (7.9,0.176)
259 83 89 91 75
0.887 0.885 0.881 0797
D 1 (7.9,0.133) (7.3,0.125) (7.0,0.128) (8.5,0.177)
105 109 110 98
0.932 0.898 0.787
F 1 (7.3, 0.108) (7.5, 0.124) (10.0, 0.205)
99 102 92
0.931 0.861
F 1 (4.5,0.084) (5.9,0.153)
110 93
0.823
R 1 (7.0,0.158)
95
H 1
I
J
K
1703142011 1704340021 1719710021
14.283 15.215 15.994
327 116 112
7.905 7.568 7.405
II II
1234 1234 1234
J K
0.936 0.885 0.774
(5.3,0.139) (5.7,0.13) (7.4,0.158)
95 95 81
0.86 0.855 0.79
(7.90.197) (7.2,0.165) (7.1,0.17)
101 100 87
0.928 0.922 0.867
(6.2,0.162) (5.3,0.117) (7.6,0.131)
247 91 87
0.879 0.836 0.721
(9.60.179) (8.5,0.154) (10.2,0.169)
310 112 108
0.915 0.902 0.84
(9.8,0.2) (9.5,0.154) (10.5,0.173)
98 95 85
0.943 0.949 0.893
(5.5,0.14) (4.3,0.1) (5.1,0.118)
101 99 89
0.915 0.953 0.873
(6.4,0.152) (4.4,0.092) (5.8,0.128)
103 101 91
0.818 0.865 0.752
(7.3,0.146) (5.1,0.124) (7.6,0.161)
92 88 78
0.922 0.809
1 (4.8,0.123) (7.1,0.157)
106 99
0.921
1 (4.2, 0.099)
90
1
Figure 3A-14. Chicago, IL MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-16
DRAFT-DO NOT QUOTE OR CITE
-------
Gary, IN MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
18-089-0006
18-089-0022
18-089-1016
18-127-0024
20
40km
b.
AIRSID#
1808900061 1808900221 1808910161 1812700241
c.
Mean
Obs
SD
80 -
60 -
40 -
20 -
0 -
1 5.455
559
8.160
,
,!
i!
in
17.0214
580
9.294
|
j!
i
I
i
I
16.447
636
8.348
1 I
II
1 1
"I
!!
'I
|i
II
13.543
196
7.571
I
I
I
Site
1234
B
D
A
B
C
D
0.666
1 (8.6, 0.205)
502
1
0.849
(5.9, 0.164)
512
0.568
(11.5, 0.220)
532
1
0.864
(7.2, 0.214)
167
0.525
(14.9, 0.265)
169
0.838
(7.9, 0.218)
177
1
Figure 3A-15. Gary, IN MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations; (c) Intersite correlation
statistics, for each data pair, the correlation coefficient, (P90, coefficient of
divergence) and number of measurements are given.
April 2002
3A-17
DRAFT-DO NOT QUOTE OR CITE
-------
Louisville, KY MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
18-019-0005
18-043-1004
21-111-0043
21-111-0044
50 km
b.
C,
AIRS ID#
Mean
Obs
SD
80 -
60 -
40 •
20 -
0 -
1801900051 1804310041 2111100431 2111100441
18.537 15.975 17.726 16.584
106 89 250 308
9.099 7.396 7.613 6.981
:
I
III 'I I l! !
1234 1234 1234 1234
Site A B C D
0.768 0.661 0.414
A 1 (7.8,0.131) (8.4,0.139) (11.2,0.23)
78 66 89
0.818 0.592
B 1 (6.3,0.148) (9.0,0.222)
59 76
0.775
C 1 (9.3, 0.143)
232
D 1
Figure 3A-16. Louisville, KY MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3 A-18
DRAFT-DO NOT QUOTE OR CITE
-------
St. Louis, MO MSA
Site A
SiteB
SiteC
Site D
AIRS Site ID
17-119-0023
17-119-1007
17-119-3007
17-163-0010
40
80 km
b.
AIRS IDS
1711900231 1711910071 1711930071 1716300101
c.
Mean
Obs
SD
50-
40-
30-
20 -
10 -
0-
20.639 17.361 15.929 17.407
115 119 117 113
8.971 7.262 6.938 7.566
I
I
I
I
I
I
I
I
,
I
I
,
I
Site
B
D
A
B
C
D
0.79 0.76703 0.67
1 (11.1, 0.145) (12.2, 0.176) (12.5, 0.192)
111 107 100
0.842 0.786
1 (6.0, 0.121) (7.9, 0.149)
107 101
0.812
1 (7.8,0.151)
105
1
Figure 3A-17. St. Louis, MO MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-19
DRAFT-DO NOT QUOTE OR CITE
-------
Baton Rouge, LA MSA
a.
Site A
Site B
SiteC
b.
r
0
AIRS IDS
Mean
Obs
SD
50 -
40 -
30 -
20 -
10 -
0 •
c. Site
20
40km
2203300021 2203300091 2212100011
15.156
218
7.109
I
I
A
[I
,,
I
I
I
I
I
ti
234
A
14.925
713
6.732
I
I
rii;
I
ll
234
B
o
1 (2.9
14.552
677
6.731
I
„
1
I
'ft
234
C
938 0
0.082) (3.2
.948
0.078)
B
C
0.96
(2.5, 0.071)
663
1
AIRS Site ID
22-033-0002
22-033-0009
22-121-0001
Figure 3A-18. Baton Rouge, LA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM2 5 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-20
DRAFT-DO NOT QUOTE OR CITE
-------
Kansas City, KS-MO, MSA
> V
wt
i~K
^X^3
m-
b
*?
4435)
b. AIRS ID#
Mean
Obs
SD
50 -
40 -
30-
20-
10 -
n -
c. Site
/
j
-jin AIRS Si
fa Site A 20-091 -C
^
-------
a.
Dallas, TX MSA
AIRS Site ID
Site A
Site B
SiteC
SiteD
SiteE
Site F
SiteG
48-085-0005
48-113-0020
48-113-0035
48-113-0050
48-113-0057
48-113-0069
48-113-0087
100km
b.
AIRS ID#
Mean
Obs
SD
40-
30-
20-
10-
0-
4808500051 4811300201 4811300351 4811300501 4811300571 4811X0691 4811300871
11.626 12.510 12.895 13.265 13.394 12.638 11.976
110 314 102 282 97 348 94
5.449 5.823 5.476 5.537 6.001 5.467 5.412
IN
:
I
i
M
III!
1234 1234 1234 1234 1234 1234 1234
c. Site A B C D E F G
0.876 0.906 0.935 0.872 0.939 0.949
A 1 (3.50.119) (4.30.108) (4.10.109) (5.6,0.13) (3.3,0.1) (2.7,0.084)
97 90 89 83 106 76
0.934 0.929 0.877 0.917 0.932
B 1 (3.40.086) (3.5,0.088) (5.30.108) (3.0,0.089) (3.6,0.089)
91 246 83 298 83
0.967 0.926 0.969 0.956
C 1 (2.0, 0.052) (2.8, 0.08) (2.2, 0.059) (3.2, 0.086)
84 79 98 72
0.918 0.974 0.948
D 1 (2.60.078) (2.2,0.056) (3.60.101)
81 273 74
0.939 0.915
E 1 (2.1,0.072) (4.9,0.114)
95 71
0.97
F 1 (3.0 0.075)
89
G 1
Figure 3A-20. Dallas, TX MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-22
DRAFT-DO NOT QUOTE OR CITE
-------
a.
Boise, ID MSA
AIRS Site ID
Site A
Site B
SiteC
Site D
16-010-0011
16-001-0017
16-027-0001
16-027-0005
b.
AIRS IDS
Mean
Obs
3D
1600100111 1600100171 1602700041 1602700051
60-
50-
40-
30-
20-
10-
9.465
257
8.029
8.857
203
7.596
8.986
238
7.657
9.943
207
8.066
II II W 111
1234
c. Site A
B
D
A
B
C
D
0.889
1 (4.4,0.155)
194
1
0.944
(3.8, 0.118)
229
0.837
(6.6, 0.184)
199
1
0.882
(6.0, 0.158)
198
0.75
(8.9, 0.231)
203
0.948
(4.1,0.109)
203
1
Figure 3A-21. Boise, ID MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM25 concentrations; (c) Intersite correlation
statistics, for each data pair, the correlation coefficient, (P90, coefficient of
divergence) and number of measurements are given.
April 2002
3A-23
DRAFT-DO NOT QUOTE OR CITE
-------
Salt Lake City, UT MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
49-011-0001
49-035-0003
49-035-0012
49-035-3006
49-035-3007
49-057-0007
30
60 km
b.
AIRS IDS
4901100011 4903500031 4903500121 4903530061 4903530071 4905700071
Mean
Obs
SD
80-
60-
40-
20-
0-
8.504 11.413 13.447 10.604 11.101 7.765
237 231 228 644 221 215
7.963 10.812 11.374 10.285 10.411 4.921
III
: i
1
'
"
I
r
lliii
j!
i
'HI
|!
, ,,:
1
c.
Site
B
D
A
B
C
D
E
0.821 0.881
1 (8.8, 0.236) (12.6, 0.281)
221 219
0.834
1 (7.6, 0.207)
214
1
0.927
(6.7, 0.204)
211
0.866
(5.4, 0.189)
205
0.92
(7.1, 0.201)
203
1
0.89
(8.5, 0.225)
208
0.828
(6.1, 0.198)
204
0.961
(5.0, 0.173)
200
0.936
(5.1, 0.167)
194
1
0.883
(3.9, 0.160)
203
0.759
(8.8, 0.213)
203
0.825
(11.8, 0.257)
199
0.883
(7.9, 0.180)
190
0.868
(8.7, 0.211)
188
Figure 3A-22. Salt Lake City, UT MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-24
DRAFT-DO NOT QUOTE OR CITE
-------
Seattle, WA MSA
\ \
•V * HP '"T Site A
-7 >£ ( Site B
, >v V Sit<= f-
) \ ^i (' Site D
-------
Portland, OR MSA
Site A
SiteB
SiteC
SiteD
40
b.
C.
AIRS Site ID
41-009-0004
41-051-0080
41-051-0244
41-067-0111
80 km
AIRS ID#
Mean
Obs
SD
125 -
100 -
75 -
50 -
25 -
0 -
4100900041 4105100201 4105102441 4106701111
6.395 9.247 8.903 7.295
338 688 669 383
4.680 7.804 5.622 6.050
I'll L
f f T T T T 1
i :
!f TTf li
i
P
1234 1234 1234 1234
Site A B C D
0.799 0.921 0.813
A 1 (6.5,0.238) (4.5,0.1935) (4.4,0.170)
319 309 313
0.747 0.894
B 1 (4.0,0.146) (4.3,0.162)
631 362
0.826
C 1 (4.6,0.18)
349
D 1
Figure 3A-24. Portland, OR MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-26
DRAFT-DO NOT QUOTE OR CITE
-------
Los Angeles-Long Beach, CA MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
06-037-0002
06-037-1103
06-037-1601
06-037-4002
06-037-9002
40
80 km
b.
AIRS ID#
0503700021 0603711031 0503716011 0603740021 0603790021
Mean
Obs
SD
100 •
75 •
50 -
25 •
0 -
21 .682
469
13.923
11
,
j
22.207
428
1 3.840
I
"
1
24.764 20.225 10.917
218 417 204
14.056 12.994 5.043
j
I
I' ll
1234
1234
c.
Site
B
D
A
B
C
D
0.828
1 (12.8,0.192)
391
1
0.763
(17.3,0.211)
196
0.88
(11.8, 0.140)
173
1
0.573
(20.2, 0.263)
379
0.752
(14.6, 0.191)
353
0.859
(11.8, 0.174)
171
1
0.276
(28.0, 0.392)
186
0.328
(26.4, 0.375)
164
0.363
(31.0,0.411)
181
0.338
(24.4, 0.356)
157
Figure 3A-25. Los Angeles-Long Beach, CA MSA. (a) Locations of sampling sites by
AIRS ID; (b) Quarterly distribution of 24-h average PM25 concentrations;
(c) Intersite correlation statistics, for each data pair, the correlation
coefficient, (P90, coefficient of divergence) and number of measurements
are given.
April 2002
DRAFT-DO NOT QUOTE OR CITE
-------
Riverside-San Bernadino, CA MSA
a.
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
06-065-1003
06-065-8001
06-071-0025
06-071-2002
06-071-9004
0 75 150km
AIRSID* 0505510031 0606580011 0507100251 0507120021 0607190041
Mean
Obs
SD
125 H
100 -
75 -
50 -
25 -
0 H
26.150
205
16.119
29.182
428
17.849
24.766 24.883
205 217
14.738 14.581
25.507
194
16.616
c.
Site
A
B
C
D
B
D
0.925 0.799 0.93 0.885
(7.0, 0.118) (15.9, 0.228) (10.7, 0.140)(10.0, 0.138)
174 173 184 164
0.721 0.888 0.847
1 (20.2, 0.252) (13.3, 0.152) (11.9, 0.147)
174 186 158
0.824 0.73
1 (15.1,0.229)(18.1,0.247)
180 162
0.91
1 (9.2,0.122)
173
Figure 3A-26. Riverside-San Bernadino, CA MSA. (a) Locations of sampling sites by
AIRS ID#; (b) Quarterly distribution of 24-h average PM2 5 concentrations;
(c) Intersite correlation statistics, for each data pair, the correlation
coefficient, (P90, coefficient of divergence) and number of measurements are
given.
April 2002
3A-28
DRAFT-DO NOT QUOTE OR CITE
-------
San Diego, CA MSA
Site A
Site B
SiteC
SiteD
b.
c.
AIRS Site ID
06-073-0001
06-073-0003
06-073-1002
06-073-1007
0 30 60 km
AIRSID#
Mean
Obs
SD
80 -
60 -
40 -
20 -
0 -
0607300011 0607300031 0607310021 0607310071
14.153 15.788 16.041 16.530
185 579 518 541
7.320 8.059 8.563 14.153
» » I I
I
I
1234 1234 1234 1234
Site A B C D
0.722 0.728 0.787
A 1 (10.1, 0.197) (10.1, 0.200) (11.1, 0.177)
144 127 139
0.818 0.699
B 1 (7.4 0.156) (11.1, 0.212)
436 441
0.693
C 1 (11.9,0.226)
385
D 1
Figure 3A-27. San Diego, CA MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average VM25 concentrations; (c) Intersite
correlation statistics, for each data pair, the correlation coefficient, (P90,
coefficient of divergence) and number of measurements are given.
April 2002
3A-29
DRAFT-DO NOT QUOTE OR CITE
-------
(a)
Columbia, SC MSA
AIRS Site ID
Site A 45-079-0019
20 40 km
(b)
AIRSID* 450790019
Mean
Obs
SD
40 H
30 -
20
10 -
1234
Figure 3A-28. Columbia, SC MSA. (a) Locations of sampling sites by AIRS ID#; (b) Quarterly
distribution of 24-h average PM10_2 5 concentrations.
April 2002
3 A-30 DRAFT-DO NOT QUOTE OR CITE
-------
Detroit, Ml MSA
AIRS Site ID
Site A
Site B
SiteC
26-163-0001
26-163-0015
26-163-0025
20
40 km
AIRS IDS
261630001 261630015 261630025
Mean
Obs
SD
80 •
60 •
40 •
20 •
0 •
-20 •
11.517
56
10.262
|
i
'!:'
19.416 7.328
58 55
15.611 7.638
\
„
1
'
1
1 „
;M|
1234 1234 1234
Site
B
A 1
B
C
0.808
0.161
148
1
0.828
0.115
167
0.77
0.158
466
1
Figure 3A-29. Detroit, MI MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_25 concentrations;
(c) Intersite correlation coefficients and number of measurements.
April 2002
3A-31
DRAFT-DO NOT QUOTE OR CITE
-------
Cleveland. OH MSA
AIRS Site ID
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
39-035-001 3
39-035-0038
39-035-0045
39-035-0060
39-035-0065
39-085-1 001
40
80 KM
AIRSIDtf
390350013 390350038 390350045 390350060 390350055 390851001
ean
Obs
SD
150 -
100 -
50 -
0 -
-50 -
26.572 18.917
109 291
18.045 11.819
'
|
II
19.032 23.072 19.190 7.411
53 51 54 51
11.1269 20.978 11.441 6.059
I
'!!
[Tf;f
ill I
kill;!*,,
1234 1234
Site A B
A 1 0.602
90
B 1
C
D
E
F
1234 1234
C D
0.715 0.691
44 43
0.67 0.735
42 42
1 0.636
44
1
1234
E
0.612
47
0.705
43
0.724
47
0.719
44
1
1234
F
0.442
43
0.426
42
0.529
43
0.216
44
0.215
45
1
Figure 3A-30. Cleveland, OH MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations;
(c) Intersite correlation coefficients and number of measurements.
April 2002
3A-32
DRAFT-DO NOT QUOTE OR CITE
-------
Steubenville, OH - Weirton, WV MSA
v\
-i\
^ }
-i / ;•
K m %
(C
b AIRS IDS 390810016
Mean 13.229
Obs 56
3D 9.682
Site A
SiteB
SiteC
SiteD
0 20 40 km
390811001 540290011 540291004
15.873 10.520 13.046
79 111 60
12.267 10.078 11.978
AIRS Site ID
39-081-0016
39-081-1001
54-029-0011
54-029-1004
100
75
50
25
-25
13.229
56
9.682
'''
l\ I
15.873
79
12.267
1
1
10.520 13.046
111 60
10.078 11.978
|
I
.I'l |l
!Trf :|H
-50 •
Site
A
B
C
D
1234 1234 1234
ABC
-I 639 0.614
37 51
-I 0.307
73
1
1234
D
0.684
55
0.633
39
0.643
54
1
Figure 3A-31. Steubenville, OH-Weirton, WV MSA. (a) Locations of sampling sites by
AIRS ID#; (b) Quarterly distribution of 24-h average PM10_25
concentrations; (c) Intersite correlation coefficients and number of
measurements.
April 2002
3A-33
DRAFT-DO NOT QUOTE OR CITE
-------
St. Louis, MO-IL MSA
Site A
SiteB
SiteC
SiteD
AIRS Site ID
17-119-0023
17-119-3007
17-163-0010
29-510-0086
40
80 km
AIRSID#
171190023 171193007 171630010 295100086
Mean
Obs
SD
100 •
75 •
50 -
25 •
0 •
-25 •
22.475 12.091 15.525 14.321
105 57 52 53
17.393 13.748 14.178 12.280
|
|
|
II
I
I
I
u
ll
ite
A
B
C
D
1234 1234 1234
ABC
-I 0.698 0.731
51 47
1 0.82
50
1
1234
D
0.679
48
0.83
51
0.837
48
1
Figure 3A-32. St. Louis, MO-IL MSA. (a) Locations of sampling sites by AIRS ID#;
(b) Quarterly distribution of 24-h average PM10_2 5 concentrations;
(c) Intersite correlation coefficients and number of measurements.
April 2002
3A-34
DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Fitz-Simons, T. S.; Mathias, S.; Rizzo, M. (2000) Analyses of 1999 PM data for the PM NAAQS review. Research
3 Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards;
4 November 17. Available: http://www.epa.gov/oar/oaqps/pm25/analyses.html [2 April, 2002].
5 Pinto, J. P.; Lefohn, A. S.; Shadwick, D. S. (2002) Aspects of the spatial variability of PM25 concentrations within
6 urban areas of the United States. Environ. Sci. Technol.: submitted.
7 Rizzo, M.; Pinto, J. P. (2001) Initial characterization of fine paniculate matter (PM2.5) collected by the National
8 Federal Reference Monitoring Network. Presented at: 94th annual conference & exhibition of the Air &
9 Waste Management Association; June, Orlando, FL. Pittsburgh, PA: Air & Waste Management Association.
10
11
12
April 2002 3A-35 DRAFT-DO NOT QUOTE OR CITE
-------
APPENDIX 3B
3 Aerosol Composition Data from the Speciation Network
4
5
6 Data from thirteen sites designed to evaluate the suitability of various aerosol sampling
7 devices for obtaining PM2 5 composition data are summarized in this appendix. Three types of
8 aerosol sampling devices were used in this study, which lasted from February 2000 through July
9 2000. A network consisting of 54 core sites across the United States has been implemented to
10 provide a consistent data set for the characterization and evaluation of trends in PM components.
11 This network has been used as a model for the deployment of a more comprehensive network,
12 consisting of approximately 250 additional sites. Data obtained from the three sampling devices
13 are shown for each site. A complete description of the data, techniques used to analyze the
14 filters, and the results of the evaluation of the performance of the sampling devices (including a
15 number of caveats regarding the data) can be found in Coutant and Stetzer (2001) and the
16 analyses of data in Coutant et al. (2001).
17 Summary statistics for concentrations of PM25 are given in Table 3B-1. Data are presented
18 for all of the sites used in the pilot study for the speciation network in Tables 3B-2 through 14.
19 Entries in the tables give the AIRS ID for each site; the number of samples (N); the mean,
20 minimum and maximum 24-h PM2 5 and component concentrations; and the minimum detection
21 limit for each constituent in the data sets for each site. Numbers given in parenthesis next to the
22 sampler indicate the POC code for identifying samplers in AIRS. Mass was determined
23 gravimetrically; anions and cations ammonium (through sulfate) were determined by ion
24 chromatography; carbonaceous species were determined by the thermal optical reflectance
25 method; and trace elements (aluminum through zirconium) were determined by X-ray
26 fluorescence spectrometry. There is a residual unknown portion ranging from <1 //m/m3 to
27 4 //g/m3, depending on the site. This residual is based on a comparison of the mass measured
28 gravimetrically with that determined by summing the contributions from measured components.
29
April 2002 3B-1 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-1. SUMMARY STATISTICS FOR PM2 5 CONCENTRATIONS DURING
FEBRUARY THROUGH JUNE 2000 OBTAINED BY COLLOCATED
FRM SAMPLERS (in
Site
N
Mean
Max
Min
Bismarck, ND (380150003)
Boston, MA (250250042)
Bronx Botanical Garden, NY
(360050083)
Chicago, IL (170310050)
Fresno, CA (060190008)
Houston, TX (482011039)
Lewis, FL (120571075)
Philadelphia, PA (421010004)
Salt Lake City, UT (490353006)
Seattle, WA (530330080)
St. Louis, MO (295100085)
60
34
62
67
86
34
59
51
35
61
68
5.97
12.53
13.87
16.39
11.12
12.24
12.50
13.93
6.52
7.37
15.15
14.30
28.70
39.00
35.80
50.00
21.90
26.70
42.50
23.70
25.00
36.80
2.50
5.10
4.70
3.10
4.00
5.90
2.87
3.70
2.50
1.90
3.10
April 2002
3B-2
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3B-2. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT BISMARCK, ND (in
~-i
to
o
o
to
OJ
td
OJ
DRAFT-
O
0
O
H
0
0
H
W
0
O
Bismarck, ND (380150003)
Parameter
PM2 5 Mass (88 101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88109)
Cadmium (88 110)
Calcium (88 111)
Carbonate Carbon (88308)
Cerium (88 117)
Cesium (88 11 8)
Chlorine (88 11 5)
Chromium (88 112)
Cobalt (881 13)
Copper (88 114)
N
23
21
20
0
21
21
25
25
25
17
14
15
23
17
7
23
0
13
12
12
9
0
16
Mean
6.35392
0.45558
0.08233
—
0.52373
1.39787
2.43559
0.23671
2.67230
0.03782
0.00645
0.00179
0.05272
0.00153
0.00403
0.05576
0.02152
0.01574
0.00231
0.00046
—
0.00081
Met One
Max
10.9262
1.18190
0.17132
—
2.03068
3.32569
4.19042
0.68209
4.85158
0.34114
0.01436
0.00394
0.09574
0.00322
0.01319
0.23228
0.07436
0.04227
0.00499
0.00138
—
0.00203
(5)
Min
3.40788
0.14508
0.02898
—
0.09675
0.75503
1.56074
0.02249
1.67502
0.00082
0.00103
0.00023
0.00336
0.00035
0.00056
0.01600
0.00460
0.00035
0.00069
0.00011
—
0.00011
MDL
0.10400
0.017
0.03
0.014
0.00800
0.012
0.146
0.146
NA
0.01088
0.01476
0.00247
0.05876
0.00199
0.0105
0.00347
0.146
0.08603
0.03689
0.00578
0.00159
0.00141
0.00135
URG (6)
N
22
21
20
10
21
21
24
24
24
14
15
14
21
21
10
22
0
10
7
4
4
2
18
Mean
4.80160
0.55753
0.03012
0.05161
0.42761
1.31732
1.46333
0.22107
1.68439
0.02811
0.00275
0.00060
0.02441
0.00137
0.00189
0.03126
0.01459
0.00620
0.00122
0.00015
0.00016
0.00042
Max
8.88982
1.17094
0.09578
0.07882
1.75630
2.01230
3.29766
0.72424
4.02190
0.27570
0.00608
0.00132
0.04242
0.00278
0.00330
0.16318
0.03447
0.01554
0.00207
0.00023
0.00019
0.00099
Min
3.04179
0.29308
0.00408
0.03138
0.06708
0.71669
0.55629
0.06229
0.79034
0.00132
0.00104
0.00019
0.00217
0.00005
0.00033
0.00899
0.00523
0.00028
0.00033
0.00005
0.00014
0.00005
MDL
0.04000
0.00700
0.01200
0.00600
0.00300
0.00500
0.05900
0.05900
NA
0.00436
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.059
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
-------
TABLE 3B-2 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT BISMARCK, ND (in
~-i
to
o
o
to
OJ
td
•^
DRAFT-
O
0
O
H
0
0
H
W
0
O
Bismarck, ND (380150003)
Parameter
Europium (88 121)
Gallium (88 124)
Gold (88 143)
Hafnium (88 127)
Indium (88131)
Iridium (88133)
Iron (88 126)
Lanthanum (88 146)
Lead (88 128)
Magnesium (88 140)
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
Sodium (88 184)
N
1
23
14
12
10
18
23
12
21
10
20
14
9
16
11
23
8
1
1
12
23
12
18
Mean
0.00300
0.00275
0.00216
0.01164
0.00443
0.00413
0.05132
0.03345
0.00382
0.01462
0.00232
0.00227
0.00117
0.00295
0.00102
0.03065
0.00118
0.00024
0.00012
0.00156
0.13816
0.00522
0.06076
Met One
Max
0.00300
0.00414
0.00601
0.02674
0.00912
0.00780
0.26884
0.05805
0.01036
0.05475
0.00990
0.00448
0.00453
0.02075
0.00287
0.15682
0.00254
0.00024
0.00012
0.00281
0.84236
0.01336
0.14392
(5)
Min
0.00300
0.00023
0.00023
0.00024
0.00093
0.00024
0.01337
0.00572
0.00012
0.00106
0.00035
0.00035
0.00025
0.00011
0.00012
0.00138
0.00011
0.00024
0.00012
0.00011
0.02970
0.00108
0.00081
MDL
0.01124
0.00331
0.00501
0.02605
0.01128
0.00594
0.00196
0.06947
0.00549
0.01841
0.00231
0.00437
0.00477
0.00125
0.00420
0.00341
0.00217
0.00617
0.00243
0.00212
0.00753
0.01048
0.05107
URG (6)
N
4
20
16
12
11
18
22
6
21
11
21
13
11
14
11
22
11
4
5
17
22
18
16
Mean
0.00178
0.00115
0.00072
0.00372
0.00197
0.00120
0.03335
0.01280
0.00228
0.00722
0.00165
0.00099
0.00111
0.00038
0.00063
0.02871
0.00037
0.00111
0.00021
0.00057
0.08587
0.00204
0.04287
Max
0.00292
0.00184
0.00207
0.00857
0.00508
0.00240
0.19338
0.02524
0.00471
0.01714
0.00631
0.00155
0.00212
0.00085
0.00141
0.13414
0.00094
0.00245
0.00047
0.00122
0.60907
0.00448
0.11010
Min
0.00080
0.00028
0.00005
0.00014
0.00013
0.00010
0.00932
0.00532
0.00071
0.00217
0.00019
0.00014
0.00005
0.00005
0.00019
0.00207
0.00005
0.00047
0.00005
0.00010
0.01615
0.00010
0.01328
MDL
0.00451
0.00133
0.00201
0.01050
0.00452
0.00238
0.00079
0.02790
0.00220
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0.00137
0.00087
0.00247
0.00097
0.00085
0.00302
0.00420
0.02050
-------
> TABLE 3B-2 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT
to
o
o
to
Parameter
Tin (88 160)
Titanium (88 161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
Zirconium (88 185)
*The blank spaces mean there are
OJ
td
BISMARCK, ND (in Mg/m3)
Bismarck, ND (380150003)
N
23
23
0
14
6
13
9
Mean
0.01887
0.00358
—
0.00837
0.00123
0.00321
0.00083
Met One
Max
0.02888
0.01910
—
0.02148
0.00264
0.01717
0.00241
(5)
Min
0.01044
0.00046
—
0.00162
0.00012
0.00083
0.00011
MDL
0.01787
0.00208
0.00150
0.01380
0.00304
0.00145
0.00359
URG (6)
N
22
22
1
15
2
18
7
Mean
0.00843
0.00228
0.00005
0.00297
0.00073
0.00206
0.00051
Max
0.01309
0.01281
0.00005
0.00537
0.00123
0.00556
0.00104
Min
0.00579
0.00037
0.00005
0.00061
0.00023
0.00014
0.00014
MDL
0.00717
0.00083
0.00060
0.00554
0.00122
0.00058
0.001
no non-zero, valid measurements.
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-3. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT BOSTON, MA (in
Boston, MA (250250042)
Parameter
PM25 Mass (88101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88 109)
Cadmium (88 110)
Calcium (881 11)
Carbonate Carbon (88308)
Cerium (88 11 7)
Cesium (88 11 8)
Chlorine (881 15)
Chromium (881 12)
Cobalt (881 13)
Copper (881 14)
Europium (88121)
Gallium (88124)
Iron (88126)
Lanthanum (88146)
Lead (88128)
N
25
21
22
3
22
22
5
5
5
14
17
19
25
25
9
25
0
16
14
24
16
1
25
0
21
25
10
25
Mean
10.6683
0.9294
0.15548
0.08220
0.94089
2.61927
3.82282
0.94296
4.76578
0.02451
0.00238
0.00094
0.02831
0.00249
0.00169
0.05016
—
0.01037
0.00376
0.06815
0.00050
0.00004
0.00222
—
0.00088
0.07623
0.01342
0.00343
Andersen
Max
24.8748
2.36412
0.59535
0.0985
4.15629
6.60445
6.18258
1.85341
8.03600
0.14572
0.00639
0.00329
0.04840
0.00895
0.00464
0.16804
—
0.03822
0.00922
1.18279
0.00348
0.00004
0.00499
—
0.00226
0.25559
0.03916
0.00636
(5)
Min
4.45285
0.07001
0.0189
0.0632
0.12428
0.45984
2.27753
0.12801
2.40554
0.00151
0.00049
0.00004
0.00602
0.00028
0.00014
0.01590
—
0.00046
0.00082
0.00004
0.00004
0.00004
0.00068
—
0.00018
0.03201
0.00014
0.00155
MDL
0.04
0.01500
0.02800
0.013
0.008
0.011
0.13400
0.134
NA
0.00436
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.13400
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
0.00451
0.00133
0.00079
0.02790
0.00220
Andersen (6)
N Mean Max Min MDL
28 10.995 25.9611 4.38247 0.04000
25 0.90881 3.499 0.06 0.015
26 0.20178 1.02994 -.00291 0.02800
1 0.16434 0.16434 0.16434 0.01300
26 0.72730 3.31728 0.09602 0.00800
26 2.6825 7.94791 0.25376 0.01100
8 4.6666 6.00087 2.53827 0.13400
8 0.9856 1.66607 0.48448 0.13400
8 5.65214 7.21195 3.34141 NA
20 0.02292 0.15922 0.00087 0.00436
15 0.00343 0.00579 0.00022 0.00592
19 0.00108 0.00235 0.00031 0.00099
28 0.03 0.05487 0.01400 0.02360
26 0 0.01 0.00087 0.00080
15 0.00187 0 0.00009 0.00421
28 0.04912 0.1648 0.01466 0.00139
0 — — — 0.13400
12 0.01043 0.02062 0 0.03450
15 0.00560 0.01054 0 0.01480
22 0.01822 0.07545 0.00111 0.00232
20 0.00070 0.00424 0.00010 0.00063
1 0.00004 0.00004 0.00004 0.00056
28 0.00244 0.00605 0.00094 0.00054
1 0.00028 0.00028 0.00028 0.00451
23 0.00075 0.00186 0.00018 0.00133
28 0.08281 0.25009 0.03017 0.00079
16 0.01184 0.02467 0 0.02790
28 0.00378 0.01053 0.00062 0.00220
URG(7)
N
27
25
25
16
25
25
9
9
9
16
21
13
26
27
10
27
0
14
16
18
15
0
27
1
22
27
12
27
Mean
10.3092
1.24094
0.12490
0.05144
0.90214
3.06590
3.46889
0.90155
4.37044
0.00595
0.00312
0.00090
0.02501
0.00279
0.00154
0.03094
—
0.01102
0.00548
0.02582
0.00023
—
0.00171
0.00043
0.00086
0.04838
0.00787
0.00337
Max
25.3771
3.4547
0.3809
0.07342
4.49158
9.01206
5.24618
1.53401
6.78019
0.01846
0.00824
0.00273
0.05663
0.00692
0.00278
0.11255
—
0.02806
0.01874
0.32562
0.00085
—
0.00433
0.00043
0.00170
0.09559
0.02496
0.00721
Min
4.29292
0.21059
0.0219
0.023
0.10802
0.45664
2.15679
0.33213
3.12194
0.00094
0.00047
0.00014
0.00424
0.00108
0.00047
0.01083
—
0.00104
0.00066
0.00066
0.00005
—
0.00047
0.00043
0.00014
0.02228
0.00259
0.00137
MDL
0.04000
0.00700
0.01200
0.00600
0.003
0.00500
0.05900
0.05900
NA
0.00436
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.05900
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
0.00451
0.00133
0.00079
0.02790
0.00220
td
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-3 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT BOSTON, MA (in
Boston, MA (250250042)
Parameter
Magnesium (88140)
Manganese (88 132)
Mercury (88142)
Molybdenum (88 134)
Nickel (881 36)
Niobium (88147)
Phosphorous (88 152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88163)
Selenium (88 154)
Silicon (88165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
Sulfur (88 169)
Tantalum (88 170)
Terbium (88 172)
Tin (88160)
Titanium (88161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
Zirconium (88185)
N
12
25
15
11
25
9
0
25
11
0
2
19
25
14
22
13
25
24
3
25
25
25
9
7
25
11
Mean
0.01609
0.00185
0.00107
0.00073
0.00284
0.00040
—
0.03758
0.00029
—
0.00032
0.00093
0.09181
0.00236
0.17809
0.00066
0.93332
0.00737
0.00166
0.00739
0.00437
0.00297
0.00189
0.00041
0.00974
0.00147
Andersen
Max
0.05037
0.00767
0.00226
0.00139
0.00810
0.00117
—
0.08191
0.00076
—
0.00045
0.00321
0.51655
0.00470
1.08304
0.00152
2.66932
0.01585
0.00299
0.01171
0.01595
0.00913
0.00416
0.00073
0.01784
0.01059
(5)
Min
0.00018
0.00004
0.00004
0.00018
0.00091
0.00009
—
0.01023
0.00004
—
0.00019
0.00004
0.01870
0.00023
0.00229
0.00004
0.17688
0.00113
0.00014
0.00131
0.00121
0.00062
0.00037
0.00009
0.00276
0.00009
MDL
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0.00251
0.00137
0.00087
0.00247
0.00097
0.00085
0.00302
0.00420
0.02050
0.00101
0.00265
0.00784
0.00302
0.00717
0.00083
0.00060
0.00554
0.00122
0.00058
0.00144
Andersen (6)
N Mean Max Min MDL
20 0.01027 0.02878 0.00071 0.00738
26 0.00175 0.00386 0.00048 0.00092
17 0.00085 0.00196 0.00004 0.00175
13 0.00072 0.00176 0.00019 0.00191
28 0.00279 0.00823 0.00086 0.00050
10 0.00069 0.00191 0.00010 0.00168
1 0 0.00067 0.00067 0.003
28 0.04 0.07416 0.01093 0.00137
10 0.00021 0.00051 0.00005 0.00087
1 0.00094 0.00094 0.00094 0.00247
3 0.00036 0.00062 0.00023 0.00097
20 0.00104 0.00255 0.00009 0.00085
28 0.09564 0.41927 0.02776 0.00302
20 0.00173 0.00416 0.00057 0.00420
24 0.15575 0.42302 0.00183 0.02050
16 0.00045 0.00125 0 0.00101
28 1.02389 2.93344 0.22349 0.00265
26 0.00707 0.01392 0.00073 0.00784
6 0.00083 0.00134 0.00042 0.00302
28 0.00785 0.01296 0.00373 0.00717
27 0.00475 0.01558 0.00109 0.00083
28 0.00323 0.01841 0.00043 0.00060
11 0.00297 0.00881 0.00004 0.00554
8 0.00033 0.00060 0.00004 0.00122
28 0.00955 0.01855 0 0
10 0 0.00165 0.00004 0.00144
URG(7)
N
13
24
9
12
27
13
0
27
6
1
3
20
27
16
25
16
27
26
3
27
26
27
8
7
27
14
Mean
0.00906
0.00102
0.00075
0.00082
0.00408
0.00043
—
0.03177
0.00022
0.00019
0.00038
0.00101
0.05312
0.00184
0.10689
0.00045
1.00354
0.007
0.00069
0.00765
0.00306
0.00376
0.00145
0.00050
0.00898
0.00072
Max
0.02392
0.00254
0.00226
0.00184
0.03146
0.00113
—
0.06259
0.00047
0.00019
0.00052
0.00315
0.13214
0.00386
0.32783
0.00094
2.77815
0.01601
0.00146
0.01267
0.00588
0.01955
0.00235
0.00094
0.01709
0.00165
Min
0.00047
0.00005
0.00010
0.00010
0.00057
0.00005
—
0.00315
0.00005
0.00019
0.00023
0.00014
0.01926
0.00014
0.00777
0.00005
0.16127
0.00033
0
0.002
0.001
0
0.00033
0.00010
0.00165
0.00005
MDL
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0.00251
0.00137
0.00087
0.00247
0.00097
0.00085
0.00302
0.00420
0.02050
0.00101
0.00265
0.008
0.00302
0.00717
0
0.00060
0.006
0.00122
0
0.001
td
-------
to
o
o
to
TABLE 3B-4. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS
BRONX BOTANICAL GARDEN, NY (in
AT
Bronx Botanical Garden, NY (360050083)
Parameter
PM2 5 Mass (88101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88 109)
Cadmium (88 110)
Calcium (881 11)
Carbonate Carbon (88308)
Cerium (88 11 7)
Cesium (88 11 8)
Chlorine (881 15)
Chromium (881 12)
Cobalt (881 13)
Copper (881 14)
Europium (88121)
Gallium (88124)
N Mean
25 12.5018
31 1.5786
31 0.16773
4 0.12953
31 1.05845
31 4.19607
6 4.20397
6 1.30671
6 5.51068
20 0.00919
19 0.00362
18 0.00113
25 0.02598
25 0.00253
10 0.00172
25 0.03839
0 —
12 0.00897
9 0.00655
16 0.00975
18 0.00034
1 0.00043
25 0.00283
0 —
18 0.00096
Andersen
Max
35.3799
5.55365
0.63440
0.16503
4.24427
13.9566
9.39840
1.87039
11.2688
0.03014
0.00667
0.00233
0.04590
0.01058
0.00312
0.09992
—
0.02442
0.01627
0.08329
0.00109
0.00043
0.00753
—
0.00176
(5)
Min
4.13505
0.11312
0.02340
0.09966
0.12851
0.60446
2.11015
0.69807
2.80822
0.00125
0.00031
0.00013
0.00535
0.00017
0.00022
0.01449
—
0.00184
0.00130
0.00017
0.00009
0.00043
0.00086
—
0.00052
MDL
0.04000
0.01500
0.02800
0.01300
0.00800
0.01100
0.13400
0.134
NA
0
0.01
0.00099
0.02360
0.00080
0.00421
0.00139
0.13400
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
0.00451
0.00133
N Mean
36 14.3136
35 1.54307
35 0.14766
5 0.13031
35 1.12762
35 3.90576
26 4.23325
26 1.31710
26 5.55035
20 0.01804
20 0.00729
23 0.00183
34 0.06317
33 0.00275
22 0
36 0.048
0 —
22 0.02991
15 0.00803
25 0.00744
13 0.00081
3 0.00175
33 0.00309
1 0.00170
33 0.00218
Met One
Max
40.0400
5.57605
0.50700
0.18103
4.30423
13.7889
8.87590
3.14339
11.0917
0.07911
0.01789
0.00412
0.09658
0.00813
0.00965
0.10880
—
0.0857
0.0307
0.0296
0.002
0.00457
0.00697
0.00170
0.00539
(6)
Min MDL
4.67955 0.10400
0.19379 0.01700
0.00990 0.03000
0.09126 0.01400
0.12779 0.00800
0.65552 0.01200
1.67583 0.14600
0.28686 0.14600
2.52934 NA
0.00136 0.01088
0.00012 0.01476
0.00012 0.00247
0.01405 0.05876
0.00035 0.00199
0.00023 0.01050
0.01593 0.00347
— 0.14600
0.00318 0.08603
0.00034 0.03689
0.00023 0.00578
0 0.00159
0 0.00141
0 0.00135
0.002 0.01124
0 0.00331
Met One (7)
N Mean
37 15.6399
36 1.51177
36 0.16675
4 0.11317
36 1.15671
36 3.80892
26 4.20562
26 1.32068
26 5.52630
26 0.01255
30 0.00724
28 0.00211
36 0.06711
34 0.00286
12 0.00369
37 0.04810
0 —
19 0.03198
21 0.02048
20 0.00729
18 0.00056
5 0.00077
36 0.00311
1 0.00012
31 0.00276
Max
43.3368
5.55544
0.53988
0.12365
4.29506
13.0896
8.81401
2.70728
10.6248
0.06945
0.02246
0.00471
0.16887
0.01289
0.00751
0.12002
—
0.08526
0.04350
0.04759
0.00131
0.00301
0.01203
0.00012
0.00529
Min
5.34188
0.15741
0.02833
0.10041
0.12589
0.58524
1.53944
0.27614
2.70700
0.00120
0.00024
0.00024
0.00636
0.00024
0.00024
0.01662
—
0.00045
0.00036
0.00012
0.00012
0.00012
0.00024
0.00012
0.00011
MDL
0.10400
0.01700
0.03000
0.01400
0.00800
0.01200
0.14600
0.14600
NA
0.01088
0.01476
0.00247
0.05876
0.00199
0.01050
0.00347
0.14600
0.08603
0.03689
0.00578
0.00159
0.00141
0.00135
0.01124
0.00331
td
oo
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
O
O
to
TABLE 3B-4 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT
BRONX BOTANICAL GARDEN, NY (in
Bronx Botanical Garden, NY (360050083)
Parameter
Iron (88126)
Lanthanum (88146)
Lead (88128)
Magnesium (88140)
Manganese (88 132)
Mercury (88142)
Molybdenum (88 134)
Nickel (88 136)
Niobium (88147)
Phosphorous (88 152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88163)
Selenium (88 154)
Silicon (88165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
Sulfur (88 169)
Tantalum (88 170)
Terbium (88 172)
Tin (88160)
Titanium (88161)
Vanadium (88 164)
N Mean
25 0.09133
15 0.01467
25 0.00419
15 0.00477
24 0.00132
14 0.00080
15 0.00089
25 0.01197
7 0.00076
1 0.00561
25 0.03778
12 0.00023
0 —
5 0.00016
19 0.00076
25 0.07495
10 0.00213
18 0.07192
20 0.00062
25 1.20653
23 0.00708
6 0.00068
25 0.00758
25 0.00407
25 0.00315
Andersen
Max
0.21643
0.03096
0.00853
0.01027
0.00360
0.00186
0.00215
0.04563
0.00187
0.00561
0.13856
0.00065
—
0.00052
0.00243
0.27206
0.00364
0.28986
0.00184
4.55104
0.01545
0.00103
0.01144
0.01006
0.00948
(5)
Min
0.04547
0.00103
0.00097
0.00109
0.00004
0.00009
0.00021
0.00392
0.00013
0.00561
0.00946
0.00004
—
0.00004
0.00013
0.02708
0.00073
0.00021
0.00009
0.20827
0.00013
0.00034
0.00154
0.00009
0.00084
MDL
0.00079
0.02790
0.00220
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0.00251
0.00137
0.00087
0
0.00097
0.00085
0.00302
0.00420
0.02050
0.00101
0.00265
0.00784
0.00302
0.00717
0.00083
0.00060
N Mean
36 0.10337
25 0.02741
36 0.00582
14 0.02012
32 0.00211
21 0.00203
15 0.00182
36 0.01722
20 0.00162
2 0.00383
36 0.03594
10 0.00047
0 —
7 0.00108
23 0.00134
36 0.09172
20 0.00358
30 0.12093
12 0.00208
36 1.34117
35 0.02138
4 0.00165
36 0.01890
36 0.00552
31 0.00383
Met One
Max
0.23509
0.08614
0.01097
0.05336
0.00596
0.00773
0.00577
0.05531
0.00451
0.00471
0.13477
0.00147
—
0.00233
0.00373
0.34880
0.01155
0.37960
0.01566
5.03203
0.05584
0.00232
0.03306
0.02499
0.01016
(6)
Min MDL
0.03960 0
0.00068 0.069
0.00103 0.01
0.00069 0.018
0.00011 0
0.00011 0.00437
0.00023 0.00477
0.00417 0.00125
0.00023 0.00420
0.00295 0.00627
0.00091 0.00341
0.00011 0.00217
— 0.01
0.00011 0.00243
0.00011 0.00212
0.00403 0.00753
0.00011 0.01048
0.00632 0.05107
0.00023 0.00251
0.09461 0.00662
0.00410 0.01954
0.00068 0.00752
0.00856 0.01787
0.00080 0.00208
0.00011 0.00150
Met One (7)
N Mean
37 0.10434
14 0.03950
36 0.00583
18 0.01857
27 0.00221
18 0.00196
18 0.00267
37 0.02253
16 0.00174
0 —
37 0.03734
8 0.00051
0 —
8 0.00110
22 0.00131
37 0.09781
24 0.00483
28 0.13348
13 0.00209
37 1.44442
35 0.02006
4 0.00156
37 0.02168
37 0.00523
36 0.00361
Max
0.26049
0.07094
0.01209
0.05926
0.00507
0.00457
0.00531
0.18701
0.00480
—
0.13799
0.00108
—
0.00241
0.00344
0.34934
0.01050
0.49236
0.01520
5.16369
0.04054
0.00216
0.03268
0.01265
0.00894
Min
0.04346
0.00381
0.00096
0.00290
0.00012
0.00012
0.00060
0.00424
0.00048
—
0.00313
0.00012
—
0.00036
0.00024
0.02746
0.00024
0.00215
0.00012
0.26699
0.00024
0.00109
0.00687
0.00132
0.00024
MDL
0.00196
0.06947
0.00549
0.01841
0.00231
0.00437
0.00477
0.00125
0.00420
0.00627
0.00341
0.00217
0.006
0.00243
0.00212
0.00753
0.01048
0.05107
0.00251
0.00662
0.01954
0.00752
0.01787
0.00208
0.00150
td
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
td
o
TABLE 3B-4 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT
BRONX BOTANICAL GARDEN, NY (in
Bronx Botanical Garden, NY (360050083)
Parameter N
Wofram(88186)
Yttrium (88183)
Zinc (88167)
Zirconium (88185)
4
9
25
8
Mean
0.00112
0.00048
0.02100
0.00059
Andersen
Max
0.00280
0.00103
0.06214
0.00104
(5)
Min
0.00030
0.00004
0.00540
0.00013
MDL
0.00554
0.00122
0.00058
0.00144
Met One (6)
N Mean Max Min MDL
17 0.00617 0.01523 0.00059 0.01380
11 0.00126 0.00271 0.00023 0.00304
36 0.02445 0.11719 0.00380 0.00145
16 0.00752 0.07212 0.00023 0.00359
Met One (7)
N
15
12
37
12
Mean
0.00390
0.00085
0.02493
0.0012
Max
0.01099
0.00182
0.10768
0.00275
Min
0.00036
0.00012
0.00421
0.00012
MDL
0.01380
0.00304
0.00145
0.004
*The blank spaces mean there are no non-zero, valid measurements.
fe
H
6
o
o
H
O
o
H
W
O
O
-------
TABLE 3B-5. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT CHICAGO, IL (in
to
o
o
to
Chicago, IL (170310050)
Andersen (5)
Parameter
PM2 5 Mass (88 101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88109)
Cadmium (88 110)
Calcium (88 111)
Carbonate Carbon (88308)
Cerium (88 117)
Cesium (88 11 8)
Chlorine (88 11 5)
Chromium (88 112)
Cobalt (881 13)
Copper (88 114)
N
33
31
31
12
31
31
33
33
33
27
21
25
33
33
20
33
0
15
17
28
30
1
33
Mean
15.6892
1.86837
0.11463
0.15903
2.1707
3.94298
4.15846
1.17651
5.33496
0.02286
0.00279
0.00123
0.03239
0.00279
0.00160
0.11438
—
0.00891
0.00915
0.02258
0.00094
0.00028
0.00312
Max
33.8161
5
0
0
8
9
6
2
9
0
0
0
0
0
0
0
0
0
0
0
0
0
17987
32829
55075
51846
26518
74101
72362
13711
17028
00614
00363
08344
00800
00410
33888
—
02859
01388
32760
00727
00028
01038
Min
4.16636
0.19935
0.01819
0.04076
0.17247
0.70557
1.40580
0.35479
1.86054
0.0016
0.001
0
0.00270
0.00046
0.00014
0.02419
—
0.00046
0.00004
0.00018
0.00004
0.00028
0.00075
MDL
0.04000
0.01500
0.02800
0.01300
0.00800
0.011
0.134
0.134
NA
0
0.01
0.00099
0.024
0
0
0
0.13400
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
URG (6)
N
35
32
32
22
32
32
35
35
35
22
23
26
35
34
21
35
0
19
12
20
27
0
35
Mean
15.4128
1.94182
0.07970
0.10708
1.90254
4.06138
3.02862
0.97663
4.00525
0.023
0
0
0.029
0
0.00142
0.09139
—
0.019
0.01
0.033
0.00099
—
0.00286
Max
33.1750
5.49383
0.37386
0.53020
8.80715
9.26054
6.39614
2.02174
7.39630
0.13164
0.00768
0.00292
0.06951
0.00782
0.004
0.28217
—
0.04235
0.01621
0.27325
0.007
—
0.00979
Min
3.45877
0.23548
0.02158
0.03710
0.07974
0.67398
0.87813
0.33258
1.21071
0.00090
0.00047
0.00005
0.00631
0.00037
0.00010
0.01956
—
0
0
0
0
—
0
MDL
0.04000
0.00700
0.01200
0.00600
0.00300
0.00500
0.05900
0.05900
NA
0.00436
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.05900
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
td
fe
H
6
o
o
H
O
O
H
W
O
O
-------
TABLE 3B-5 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT CHICAGO, IL (in
to
o
o
to
Chicago, IL (170310050)
Andersen (5)
Parameter
Indium (88131)
Indium (88133)
Iron (88 126)
Lanthanum (88 146)
Lead (88 128)
Magnesium (88 140)
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Phosphorous (88152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
Sulfur (88 169)
Tantalum (88 170)
N
19
18
33
18
33
22
33
17
16
30
13
0
33
17
2
2
32
33
23
13
27
33
30
Mean
0.00155
0.00129
0.16423
0.01258
0.00864
0.01423
0.00769
0.00068
0.00103
0.00136
0.00102
—
0.09270
0.00040
0.00027
0.00055
0.00146
0.13614
0.00197
0.05172
0.00106
1.40800
0.00659
0
0
0
0
Max
00499
00298
47993
02731
0.02078
0.08985
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
03205
00230
00235
01287
00232
—
58173
00111
00032
00065
00414
66607
00464
14020
00747
85499
01553
Min
0.00009
0.00014
0.03366
0.00037
0.00065
0.00068
0.00056
0.00004
0.00014
0.00018
0.00042
—
0.01370
0.00004
0.00023
0.00046
0.00014
0.02395
0.00051
0.00698
0.00004
0.30327
0.00005
MDL
0.00452
0.00238
0.00079
0.02790
0.00220
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0
0.00137
0.00087
0.00247
0.00097
0.00085
0.00302
0.00420
0.02050
0.00101
0.00265
0.00784
URG (6)
N
18
18
35
23
35
25
35
14
18
32
11
0
35
13
1
3
32
35
27
18
23
35
31
Mean
0.00185
0.00112
0.14520
0.01227
0.00818
0.00866
0.00656
0.00096
0.00121
0.00175
0.00053
—
0.08544
0.00046
0.00061
0.00031
0.00165
0.11101
0.00186
0.05275
0.00104
1.47025
0.00586
Max
0.00410
0.00414
0.45138
0.03647
0.01871
0.03057
0.02807
0.00212
0.00565
0.01361
0.00094
—
0.56626
0.00127
0.00061
0
0.005
0.48048
0.004
0.16799
0.00608
4.02403
0.01234
Min
0
0
0.03
0
0
0
0.00057
0.00019
0.00014
0.00014
0.00010
—
0.01272
0.00005
0.00061
0
0.00014
0.02072
0.00010
0.00434
0.00005
0.29744
0.00005
MDL
0.00452
0.00238
0.0008
0.0279
0.00220
0.00738
0.0009
0.00175
0.00191
0.00050
0.00168
0.00251
0.00137
0.00087
0.00247
0.001
0.0009
0.00302
0.0042
0.02050
0.00101
0.00265
0.00784
td
to
fe
H
6
o
o
H
O
O
H
W
O
O
-------
TABLE 3B-5 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT CHICAGO, IL (in
to
o
o
to
Chicago, IL (170310050)
Andersen (5)
Parameter
Terbium (88 172)
Tin (88 160)
Titanium (88 161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
Zirconium (88 185)
N
12
33
32
15
3
11
33
18
Mean
0.00217
0.00963
0.00456
0.00090
0.00256
0.00060
0.04496
0.00061
0
0
0
0
0
0
0
0
Max
00538
01947
01272
00302
00322
00156
14491
00139
Min
0.00005
0.00341
0.00041
0.00004
0.00218
0.00004
0.00238
0.00004
MDL
0.00302
0.00717
0.00083
0.00060
0.00554
0.00122
0.00058
0.00144
URG (6)
N
10
35
34
18
4
12
35
19
Mean
0.00139
0.00985
0.00423
0.00080
0.00099
0.00029
0.04233
0.00075
Max
0.00396
0.01611
0.00979
0.00348
0.00283
0.00057
0.13774
0.00179
Min
0
0
0.00108
0.00005
0.00014
0.00005
0.00188
0
MDL
0.00302
0.00717
0.0008
0.0006
0.00554
0.00122
0.00058
0.00144
td
fe
H
6
o
o
H
O
O
H
W
O
*The blank spaces mean there are no non-zero, valid measurements.
O
-------
>
to
o
o
to
TABLE 3B-6. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT FRESNO, CA (in
Fresno, CA (060190008)
Parameter
PM2 5 Mass (88 101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88109)
Cadmium (88 110)
Calcium (88 111)
Carbonate Carbon (88308)
Cerium (88 117)
Cesium (88 11 8)
Chlorine (88 11 5)
Chromium (88 112)
Cobalt (881 13)
Copper (88 114)
Europium (88 121)
Gallium (88 124)
N
25
20
20
7
20
20
24
24
24
25
16
12
24
23
8
25
0
18
14
18
13
0
24
2
23
Mean
10.4804
0.73724
0.25999
0.20847
1.41616
1.77563
4.71732
0.51751
5.23483
0.03524
0.00731
0.00147
0.06186
0.00238
0.00604
0.05638
-
0.03076
0.01431
0.01254
0.00064
-
0.00309
0.00431
0.00180
Met One (5)
Max
16.7861
1.64233
0.62721
0.48186
3.58681
2.78369
8.20203
0.95420
8.77901
0.09154
0.01520
0.00357
0.09844
0.00460
0.01011
0.09913
-
0.06598
0.06182
0.08483
0.00151
-
0.01162
0.00461
0.00466
Min
6.08247
0.21479
0.10849
0.05730
0.36860
0.60374
2.96153
0.12369
3.10442
0.00596
0.00011
0.00035
0.01265
0.00024
0.00150
0.02704
-
0.00150
0.00011
0.00139
0.00011
-
0.00080
0.00401
0.00023
MDL
0.10400
0.01700
0.03000
0.01400
0.00800
0.01200
0.14600
0.14600
NA
0.01088
0.01476
0.0025
0.05876
0.002
0.01050
0.00347
0.14600
0.08603
0.03689
0.00578
0.00159
0.00141
0.00135
0.01124
0.00331
Andersen (6)
N
25
20
20
4
20
20
24
24
24
25
18
16
25
25
11
25
0
15
12
22
20
2
25
2
20
Mean
9.11039
0.59866
0.21312
0.14457
1.27893
1.55442
4.45785
0.52888
4.98674
0.03758
0.00209
0.00066
0.02736
0.002
0.001
0.0546
-
0.0107
0.005
0.009
0
0.00029
0.00351
0.00163
0.00091
Max
18.7161
1.23704
0.40863
0.37056
3.04898
2.46932
7.47617
0.99233
8.11572
0.08153
0.00750
0.00105
0.06863
0.00579
0.00287
0.08277
-
0.0417
0.0141
0.0674
0.004
0
0.013
0.002
0.003
Min
5.11270
0.16523
0.07708
0.06370
0.51775
0.53107
2.78310
0.10721
2.95225
0.00173
0.00026
0.00010
0.00179
0.00085
0.00023
0.02980
-
0.00029
0.00005
0.00014
0
0
0
0.0012
0.00010
MDL
0.04000
0.01500
0.02800
0.01300
0.00800
0.01100
0.13400
0.13400
NA
0.00436
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.13400
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
0.00451
0.00133
td
j^.
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-6 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT FRESNO, CA (in
Fresno, CA (060190008)
Parameter
Iron (88 126)
Lanthanum (88 146)
Lead (88 128)
Magnesium (88 140)
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Phosphorous (88152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
Sulfur (88 169)
Tantalum (88 170)
Terbium (88 172)
Tin (88 160)
Zinc (88167)
Zirconium (88 185)
N
25
14
25
13
23
10
11
24
7
0
25
8
0
5
19
25
13
20
8
25
23
0
25
23
11
Met One (5) Andersen (6)
Mean Max Min MDL N Mean Max Min MDL
0.08889 0.14245 0.04555 0.00196 25 0.08660 0.14179 0.03598 0.00079
0.03589 0.08080 0.01508 0.06947 14 0.01508 0.03212 0.00256 0.02790
0.00395 0.00795 0.00035 0.00549 25 0.00285 0.00991 0.00033 0.00220
0.01738 0.05313 0.00184 0.01841 17 0.01300 0.05139 0.00014 0.00738
0.00205 0.00441 0.00023 0.00231 25 0.00212 0.00430 0.00020 0.00092
0.00132 0.00322 0.00023 0.00437 14 0.00106 0.00192 0.00014 0.00175
0.00194 0.00391 0.00011 0.00477 10 0.00098 0.00223 0.00010 0.00191
0.00933 0.02900 0.00011 0.00125 19 0.00061 0.00140 0.00010 0.00050
0.00085 0.00150 0.00011 0.00420 11 0.00078 0.00172 0.00015 0.00168
— — — 0.0063 0 — — — 0.00627
0.07233 0.41635 0.02418 0.00341 25 0.07447 0.41798 0.02572 0.00137
0.00127 0.00220 0.00011 0.00217 6 0.00022 0.00067 0.00005 0.00087
— — — 0.00617 1 0.00020 0.00020 0.00020 0.00247
0.00067 0.00115 0.00023 0.00243 4 0.00026 0.00047 0.00014 0.00097
0.00191 0.00587 0.00011 0.00212 20 0.00169 0.00407 0.00030 0.00085
0.18161 0.32125 0.09166 0.00753 25 0.17293 0.27442 0.05783 0.00302
0.00419 0.00808 0.00011 0.01048 14 0.00212 0.00577 0.00005 0.00420
0.14560 0.43413 0.02406 0.05107 20 0.10596 0.45392 0.00715 0.02050
0.00240 0.00518 0.00057 0.00251 15 0.00093 0.00619 0.00010 0.00101
0.58123 1.22825 0.23819 0.00662 25 0.54945 1.22229 0.21920 0.00265
0.01896 0.03739 0.00413 0.01954 24 0.00822 0.01775 0.00029 0.00784
— — — 0.00752 7 0.00134 0.00263 0.00033 0.00302
0.02086 0.03423 0.01152 0.01787 25 0.00890 0.01476 0.00371 0.00717
0.00615 0.06497 0.00011 0.00145 25 0.02414 0.08440 0.00733 0.00058
0.00127 0.00290 0.00011 0.00359 11 0.00100 0.00371 0 0.00144
td
fe
H
6
o
o
H
O
O
H
W
O
O
*The blank spaces mean there are no non-zero, valid measurements.
-------
TABLE 3B-7. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT HOUSTON, TX (in
to
o
o
to
Houston, TX (482011039)
Parameter
PM2 5 Mass (88 101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88109)
Cadmium (88 110)
Calcium (88 111)
Carbonate Carbon (88308)
Cerium (88 117)
Cesium (88 11 8)
Chlorine (88 11 5)
Chromium (88 112)
Cobalt (881 13)
Copper (88 114)
N
25
23
23
14
23
23
20
20
20
18
15
20
25
25
9
25
0
12
12
23
21
0
25
Mean
14.1882
0.80390
0.56265
0.15733
0.84851
3.84944
2.45022
0.39000
2.84022
0.21171
0.00334
0.00103
0.02703
0.00428
0.00224
0.10541
—
0.00861
0.00522
0.15774
0.00067
—
0.00228
Andersen (5)
Max
23.8011
3.32458
1.78638
0.33064
2.50829
10.6928
3.65958
0.7091
4.05657
1.22376
0.00619
0.00266
0.04721
0.01236
0.00665
0.41857
—
0.01814
0.01497
1.16485
0.00160
—
0.01601
7
0
0
0
0
1
1
0
1
Min
31156
06796
07091
10210
29509
06609
56175
04994
.6586
0.001
0.001
0
0
0
0
0
0
0
0
0
0
00449
00045
00022
03970
—
00221
00009
00077
00014
—
00045
MDL
0.04000
0.01500
0.02800
0.01300
0.00800
0.01100
0.13400
0.13400
NA
0.004
0.006
0
0.0236
0
0.004
0.001
0.13400
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
URG (6)
N
24
22
22
20
22
22
19
19
19
20
16
20
23
24
11
24
0
11
11
20
19
1
24
Mean
11.9393
1.26974
0.34266
0.12180
0.74211
3.89597
1.84457
0.31709
2.16165
0.13097
0.00292
0.00095
0.02388
0.00364
0.001
0.0667
—
0.01
0.005
0.0988
0
0
0.002
Max
17.6740
2.67911
1.09133
0.38460
2.95220
7.13703
3.74028
0.55817
4.24789
0.92532
0.00852
0.00221
0.03965
0.01304
0.00475
0.26045
—
0.0282
0.009
0.60174
0
0.00028
0.02350
Min
5.16860
0.33023
0.05162
0.04691
0.24085
1.15058
0.72794
0.07787
0.85714
0.00023
0.00038
0.00010
0.00400
0.00010
0.00019
0.02006
—
0.00170
0.00023
0.00014
0
0.0003
0.0002
MDL
0.04000
0.00700
0.01200
0.00600
0.00300
0.00500
0.05900
0.05900
NA
0.00436
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.05900
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
td
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-7 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT HOUSTON, TX (in
Houston, TX (482011039)
Andersen (5)
Parameter
Europium (88 121)
Gallium (88 124)
Gold (88 143)
Hafnium (88 127)
Indium (88131)
Indium (88133)
Iron (88 126)
Lanthanum (88 146)
Lead (88 128)
Magnesium (88 140)
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Phosphorous (88152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
N
0
22
9
9
10
19
25
17
25
11
25
16
10
25
14
0
25
12
0
2
20
25
13
Mean
—
0.00113
0.00074
0.00536
0.00168
0.00124
0.12309
0.01132
0.00283
0.02578
0.00330
0.00086
0.00102
0.00155
0.00061
—
0.12207
0.00049
—
0.00012
0.00078
0.36800
0.00225
0
Max
—
00193
0.00187
0
0
0
01202
00379
00330
0.71419
0.02234
0
0
0
0
0
0
0
0
0
0
00703
06027
01214
00201
00275
00306
00219
—
29581
00114
—
00013
0.00161
2
0
30769
00504
Min
—
0.00031
0.00004
0.00033
0.00039
0.00005
0.02203
0.00060
0.00037
0.00975
0.00041
0.00022
0.00009
0.00030
0.00009
—
0.03480
0.00010
—
0.00010
0.00009
0.03828
0.00041
MDL
0.005
0.00133
0.00201
0.01050
0.00452
0.00238
0.00079
0.02790
0.00220
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0.003
0.00137
0.00087
0.002
0.00097
0.00085
0.00302
0.00420
URG (6)
N
0
19
9
14
9
15
24
10
24
10
20
10
15
24
12
0
24
13
0
1
18
24
16
Mean
—
0.001
0.001
0.003
0.002
0.00145
0.08674
0.01324
0.00277
0.02440
0.00233
0.00071
0.00122
0.00239
0.00089
—
0.10719
0.00054
—
0.00014
0.00060
0.27270
0.00188
Max
—
0.00198
0.00268
0.00810
0.004
0.003
0.47667
0.02119
0.00494
0.05015
0.00945
0.00198
0.00410
0.00866
0.00198
—
0.34553
0.002
—
0
0.002
1.72228
0.004
Min
—
0.0001
0.0001
0.0001
0.00066
0.00023
0.01385
0.00250
0.00071
0.00010
0.0001
0.0001
0
0.0002
0.00010
—
0.03123
0
—
0.0001
0
0.02576
0
MDL
0.00451
0.00133
0.00201
0.0105
0.00452
0.00238
0.00079
0.0279
0.00220
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0.00251
0.00137
0.00087
0.00247
0.00097
0.00085
0.00302
0.00420
td
-------
TABLE 3B-7 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT HOUSTON, TX (in
to
o
o
to
Houston, TX (482011039)
Parameter
Terbium (88 172)
Tin (88 160)
Titanium (88 161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
Zirconium (88 185)
N
3
25
25
24
10
5
25
11
Mean
0.00107
0.00817
0.01120
0.00311
0.00282
0.00043
0.00657
0.00072
Andersen (5)
Max
0.00135
0.01260
0.06742
0.00814
0.00591
0.00086
0.02032
0.00196
0
0
0
0
0
0
0
0
Min
00055
00372
00197
00074
00020
00009
00060
00014
MDL
0.00302
0.00717
0.00083
0.00060
0.00554
0.00122
0.00058
0.00144
URG (6)
N
4
24
24
23
13
7
24
10
Mean
0.00124
0.00737
0.00785
0.00313
0.00281
0.00040
0.00560
0
Max
0.00307
0.01304
0.04432
0.00815
0.00669
0.00085
0.01926
0.00113
Min
0.0001
0.00217
0.00141
0.00014
0.00066
0.00005
0.00033
0.00010
MDL
0.00302
0.00717
0.0008
0.0006
0.00554
0.00122
0.0006
0.00144
td
oo
fe
H
6
o
o
H
O
O
H
W
O
*The blank spaces mean there are no non-zero, valid measurements.
O
-------
TABLE 3B-8. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT LEWIS, FL (in
2_
to
O
o
to
OJ
td
VO
O
£
H
6
o
o
H
0
0
H
W
0
O
p
Lewis, FL (120571075)
Parameter
PM2 5 Mass (88101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88 109)
Cadmium (88 110)
Calcium (881 11)
Carbonate Carbon (88308)
Cerium (88 11 7)
Cesium (88 11 8)
Chlorine (881 15)
Chromium (881 12)
Cobalt (881 13)
Copper (881 14)
Europium (88121)
Iron (88126)
Lanthanum (88146)
Gallium (88124)
N
34
35
34
11
35
35
37
37
37
30
26
26
34
33
20
34
0
21
15
21
14
3
30
1
34
16
31
Mean
14.6122
1.10695
0.31849
0.18475
0.56896
4.43481
3.46126
0.55103
4.01229
0.03213
0.00484
0.00247
0.05813
0.00317
0.00322
0.07670
0.02962
0.01518
0.03747
0.00076
0.00045
0.00221
0.00023
0.05842
0.03758
0.00261
MetOne
Max
35.7215
3.45870
1.14027
0.9625
1.48237
12.1344
8.02624
1.27681
8.56808
0.23294
0.01885
0.00569
0.12715
0.00817
0.01160
0.13849
0.07865
0.03517
0.29395
0.00156
0.00101
0.01852
0.00023
0.17794
0.07532
0.00605
(5)
Min
6.96379
0.03176
0.11746
0.04954
0.03221
1.71479
1.42506
0.14487
1.94724
0.00168
0.0003
0.00023
0.00716
0.00032
0.00023
0.03215
0.00089
0.00078
0.00134
0.00011
0.00011
0.00011
0.00023
0.02057
0.00011
0.00045
MDL
0.10400
0.01700
0.03000
0.01400
0.00800
0.01200
0.146
0.146
NA
0.01088
0.01476
0.00247
0.05876
0.00199
0.0105
0.00347
0.14600
0.08603
0.03689
0.00578
0.00159
0.00141
0.00135
0.01124
0.00196
0.06947
0.00331
URG (6)
N
32
33
33
27
33
33
35
35
35
23
16
24
32
32
16
32
0
15
13
18
20
3
32
1
32
23
29
Mean
12.7384
1.68153
0.15983
0.10138
0.50700
4.47554
2.50716
0.41304
2.92020
0.02566
0.00348
0.00162
0.02719
0.00342
0.00221
0.05308
0.00855
0.00447
0.01529
0.0003
0.0002
0.00172
0.0007
0.04877
0.01039
0.00105
Max
32.0973
3.80897
0.45793
0.91296
1.46923
9.88074
6.87372
0.72688
7.04766
0.22656
0.00896
0.00527
0.07907
0.00947
0.00457
0.08769
0.01756
0.01410
0.13924
0.00137
0.00033
0.01761
0.0007
0.16403
0.02383
0.00311
Min
6.50163
0.53032
0.02519
0.03434
0.15998
1.86214
1.27824
0.16214
1.45098
0.00259
0.00099
0.00028
0.00099
0.00071
0.00047
0.019
0.00104
0.00010
0.00047
0.00005
0.00005
0.00028
0.00066
0.01763
0.00080
0.00023
MDL
0.04000
0.00700
0.01200
0.00600
0.00300
0.00500
0.05900
0.05900
NA
0.00436
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.059
0.0345
0.0148
0.00232
0.00063
0.00056
0.00054
0.00451
0.00079
0.02790
0.00133
-------
TABLE 3B-8 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT LEWIS, FL (in
2_
to
O
o
to
OJ
td
to
o
DRAFT-
O
O
O
H
0
0
H
W
0
w
Lewis, FL (120571075)
Parameter
Lead (88 128)
Magnesium (88 140)
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Phosphorous (88152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
Sulfur (88 169)
Tantalum (88 170)
Terbium (88 172)
Tin (88 160)
Zinc (88167)
N
34
14
28
23
15
31
16
0
34
13
3
7
22
34
22
31
13
34
34
3
34
33
Mean
0.00557
0.02254
0.00194
0.00206
0.00210
0.00554
0.00136
—
0.07957
0.00079
0.00052
0.00067
0.00110
0.14964
0.00449
0.17202
0.00189
1.57397
0.02010
0.00052
0.01868
0.00568
MetOne
Max
0.02700
0.04844
0.00470
0.00513
0.00537
0.05896
0.00274
—
0.89347
0.00246
0.00112
0.00145
0.00268
0.58150
0.00902
0.46907
0.01060
4.30113
0.04911
0.00101
0.03130
0.01264
(5)
Min
0.00122
0.00805
0.00011
0.00011
0.00023
0.00011
0.00011
—
0.01909
0.00011
0.00011
0.00011
0.00011
0.05025
0.00045
0.00903
0.00022
0.64698
0.00136
0.00011
0.00291
0.00033
MDL
0.00549
0.01841
0.00231
0.00437
0.00477
0.00125
0.00420
0.00617
0.00341
0.00217
0.00617
0.00243
0.00212
0.00753
0.01048
0.05107
0.00251
0.00662
0.01954
0.00752
0.01787
0.00145
URG (6)
N
32
16
31
14
18
30
14
0
32
14
0
2
24
32
24
29
16
32
31
4
32
32
Mean
0.00438
0.00912
0.00139
0.00053
0.00088
0.00203
0.00062
—
0.08082
0.00038
—
0.00014
0.00058
0.12164
0.00184
0.11949
0.00140
1.45578
0.00867
0.00065
0.00838
0.00625
Max
0.02519
0.02081
0.00325
0.00151
0.00193
0.01554
0.00151
—
0.89697
0.00108
—
0.00014
0.00162
0.49329
0.00444
0.41652
0.01224
3.33434
0.01990
0.00132
0.01551
0.01263
Min
0.00080
0.00160
0.00028
0.00005
0.00005
0.00028
0.00005
—
0.02027
0.00005
—
0.00014
0.00005
0.03804
0.00014
0.01269
0.00019
0.67286
0.00146
0.00005
0.00198
0.00165
MDL
0.00220
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0.00251
0.00137
0.00087
0.00247
0.00097
0.00085
0.00302
0.00420
0.02050
0.00101
0.00265
0.00784
0.00302
0.00717
0.00058
O
*The blank spaces mean there are no non-zero, valid measurements.
-------
TABLE 3B-9. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT PHILADELPHIA, PA (in
to
o
o
to
Philadelphia, PA (421010004)
Parameter
PM2 5 Mass (88101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88 109)
Cadmium (88 110)
Calcium (881 11)
Carbonate Carbon (88308)
Cerium (88 11 7)
Cesium (88 11 8)
Chlorine (881 15)
Chromium (881 12)
Cobalt (881 13)
Copper (881 14)
Europium (88121)
Gallium (88124)
N
37
38
37
7
38
38
37
37
37
30
26
30
37
37
18
37
0
18
17
27
28
0
37
2
29
Mean
14.7207
1.60087
0.18607
0.18764
1.38838
4.34017
4.17646
0.83466
5.01112
0.01787
0.00354
0.00101
0.02638
0.00340
0.00180
0.05694
—
0.01002
0.00563
0.00771
0.00107
—
0.00453
0.00044
0.00091
Andersen (5)
Max
43.1932
4.62702
0.73099
0.37397
3.69438
13.8852
10.5209
1.89372
11.2164
0.07417
0.00813
0.00241
0.04363
0.00828
0.00483
0.15631
—
0.01922
0.01665
0.06483
0.00386
—
0.01252
0.00052
0.00228
Min
3.75162
0.20477
0.03102
0.06598
0.21257
0.9297
1.41139
0.18759
1.59898
0.00163
0.00049
0.0001
0.00443
0.00070
0
0.01194
—
0.00071
0.00010
0.00010
0.00004
—
0.00034
0.00035
0.00010
MDL
0.04000
0.01500
0.02800
0.01300
0.00800
0.011
0.134
0.13400
NA
0.0044
0.0059
0.001
0.0236
0.00080
0.0042
0.0014
0.134
0.03450
0.0148
0.0023
0.001
0.001
0.001
0.0045
0.0013
URG (6)
N
37
38
38
21
38
38
37
37
37
21
21
29
37
37
20
37
0
16
17
16
27
2
37
0
31
Mean
13.3653
1.89732
0.07003
0.07957
1.23693
4.43424
3.20367
0.66304
3.86671
0.0101
0.003
0.00111
0.02628
0.00334
0.002
0.0342
—
0.01305
0.00648
0.00495
0.00055
0.00016
0.003
—
0
Max
39.0898
5.62864
0.19235
0.27421
4.35089
14.2670
8.35341
1.53473
8.75990
0.03047
0.00725
0.003
0.047
0.00810
0.00414
0.10905
—
0.02825
0.01738
0.01695
0.00160
0.00028
0.00579
—
0.00212
Min
3.66728
0.36506
0.01947
0.03753
0.14346
0.86455
0.97135
0.26893
1.24027
0.00010
0.00014
0.00005
0.00212
0.0003
0.00005
0.00946
—
0.00179
0.00080
0.00033
0.00005
0.00005
0.00080
—
0.00005
MDL
0.04000
0.00700
0.01200
0.00600
0.00300
0.00500
0.05900
0.05900
NA
0.00436
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.05900
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
0.00451
0.00133
td
to
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-9 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS
PHILADELPHIA, PA (in
AT
Philadelphia, PA (421010004)
Parameter
Iron (88 126)
Lanthanum (88 146)
Lead (88 128)
Magnesium (88 140)
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Phosphorous (88152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
Sulfur (88 169)
Tantalum (88 170)
Terbium (88 172)
Tin (88 160)
N
37
18
37
18
37
19
22
35
18
0
37
15
0
6
31
37
23
23
26
37
35
4
37
Andersen (5) URG (6)
Mean Max Min MDL N Mean Max Min MDL
0.10270 0.22846 0.02972 0.00079 37 0.0662 0.15637 0.02608 0.00079
0.00968 0.02769 0.00023 0.02790 17 0.00693 0.0167 0.00037 0.02790
0.00556 0.01189 0.00161 0.00220 37 0.00513 0.0106 0.00165 0.00220
0.00773 0.01847 0.00053 0.00738 17 0.00738 0.0208 0.00010 0.00738
0.00268 0.00700 0.00029 0.00092 35 0.00171 0.00372 0.00033 0.00092
0.00092 0.00232 0.00005 0.00175 20 0.00084 0.00155 0.00005 0.00175
0.00120 0.00289 0.00004 0.00191 16 0.00098 0.00325 0.00005 0.00191
0.00441 0.02189 0.00005 0.00050 37 0.00413 0.01714 0.00005 0.00050
0.00078 0.00167 0.00010 0.00168 24 0 0.002 0.00005 0.00168
— — — 0.0025 0 — — — 0.00251
0.05987 0.29381 0.00959 0.00137 37 0.0506 0.30154 0.00588 0.00137
0.00037 0.00085 0.00005 0.00087 12 0.00050 0.001 0.0001 0.00087
— — — 0.00247 1 0.00010 0 0.0001 0.00247
0.00033 0.00057 0.00010 0.00097 12 0.00035 0.00071 0 0.001
0.00113 0.00330 0.00010 0.00085 32 0.00108 0.003 0.00005 0.0009
0.11766 0.41821 0.02285 0.00302 37 0.07632 0.32847 0.01992 0.00302
0.00186 0.00503 0.00015 0.00420 25 0.00201 0.00424 0.00028 0.00420
0.06349 0.19720 0.00069 0.02050 27 0.04800 0.17957 0.00410 0.02050
0.00091 0.00515 0.00010 0.00101 22 0.00097 0.00447 0.00010 0.00101
1.50616 5.23630 0.32794 0.00265 37 1.49876 5.21337 0.30435 0.00265
0.00715 0.01651 0.00019 0.00784 35 0.00765 0.01624 0.00014 0.00784
0.00087 0.00266 0.00005 0.00302 4 0.00072 0.00108 0.00019 0.00302
0.00913 0.01527 0.00189 0.00717 37 0.00864 0.01450 0.00146 0.00717
td
to
to
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-9 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT
PHILADELPHIA, PA (in
Philadelphia, PA (421010004)
Parameter
Titanium (88 161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
N
37
30
10
11
37
Mean
0.00590
0.00358
0.00210
0.00039
0.01588
Andersen (5)
Max
0.01489
0.01202
0.00355
0.00123
0.04560
0
0
0
0
0
Min
00124
00005
00043
00005
00190
MDL
0.00083
0.00060
0.00554
0.00122
0.00058
URG (6)
N
36
29
6
12
37
Mean
0.00373
0.00354
0.00254
0.00036
0.01372
Max
0.01191
0.01140
0.00452
0.00122
0.04045
Min
0.00099
0.00019
0.00058
0.00010
0.00113
MDL
0.00083
0.00060
0.00554
0.00122
0.0006
td
to
fe
H
6
o
o
H
O
O
H
W
O
*The blank spaces mean there are no non-zero, valid measurements.
O
-------
to
o
o
to
TABLE 3B-10. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT PHOENIX, AZ (in
Phoenix, AZ (040139997)
Parameter
PM2 5 Mass (88101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88 109)
Cadmium (88 110)
Calcium (881 11)
Carbonate Carbon (88308)
Cerium (88 11 7)
Cesium (88 11 8)
Chlorine (881 15)
Chromium (881 12)
Cobalt (881 13)
Copper (881 14)
Europium (88121)
Gallium (88124)
Gold (88143)
Lanthanum (88146)
Lead (88128)
N
31
33
33
28
33
33
33
33
33
31
17
20
31
31
13
31
0
17
14
31
20
1
31
3
29
13
16
31
Mean
7.90504
0.40373
0.14516
0.08571
0.44349
1.22486
3.32080
0.61949
3.94028
0.12266
0.00275
0.00092
0.02565
0.00392
0.00179
0.13114
—
0.00735
0.00607
0.06039
0.00031
0.00019
0.00256
0.00024
0.00094
0.00084
0.01360
0.00312
URG(5)
Max
14.0493
0.73590
0.37517
0.22077
1.53715
2.16162
7.27046
1.37649
8.64695
0.50254
0.00501
0.00268
0.06113
0.00825
0.00433
0.33690
—
0.02552
0.01577
0.22815
0.00080
0.00019
0.00725
0.00033
0.00222
0.00226
0.02868
0.00636
Min
3.75000
0.14333
0.05387
0.0381
0.18007
0.46270
1.44303
0.14996
1.59299
0.00833
0.00019
0.00010
0.00085
0.00108
0.00010
0.02783
—
0.00028
0.00066
0.00160
0.00005
0.00019
0.00043
0.00014
0.00005
0.00010
0.00071
0.00160
MDL
0.04000
0.00700
0.01200
0.00600
0.003
0.005
0.059
0.05900
NA
0.004
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.05900
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
0.00451
0.00133
0.00201
0.02790
0.00220
URG (6) MetOne (7)
N Mean Max Min MDL N Mean Max Min MDL
31 7.92975 13.054 3.54388 0.04000 28 9.77348 15.1378 5.53264 0.10400
31 0.3968 0.7288 0.14380 0.00700 30 0.39832 0.76107 0.12791 0.01700
31 0.1347 0.35604 0.03440 0.01200 30 0.17580 0.39862 0.05086 0.03000
25 0.088 0.2091 0.03403 0.00600 8 0.14045 0.19476 0.07246 0.01400
31 0.50471 2.1011 0.16563 0.00300 30 0.60367 2.02319 0.20640 0.00800
31 1.19059 2.0969 0.4645 0.00500 30 1.26266 2.27230 0.57097 0.01200
30 3.49197 6.43129 1.74887 0.05900 25 4.51393 8.86971 2.67486 0.14600
30 0.6383 1.3136 0.12856 0.05900 25 0.74349 1.66706 0.23419 0.14600
30 4.1302 7.7449 1.87742 NA 25 5.25742 10.2845 3.12239 NA
31 0.11959 0.5118 0.01 0.00436 28 0.17061 0.59217 0.00815 0.01088
18 0.00317 0.01 0 0.00592 14 0.00492 0.01123 0.00011 0.01476
20 0.00079 0 0.00005 0 20 0.00158 0.00559 0.00023 0.00247
30 0.02804 0.049 0.00938 0.02360 27 0.06177 0.12311 0.00034 0.05876
31 0.00407 0.00811 0 0.00080 28 0.00322 0.00638 0.00034 0.00199
15 0.00124 0.00311 0 0.00421 14 0.00527 0.01050 0.00034 0.01050
31 0.13075 0.36095 0.031 0.00139 28 0.18478 0.42720 0.03091 0.00347
0 — — — 0.05900 0 — — — 0.14600
18 0.01519 0.03043 0 0.03450 14 0.02084 0.04110 0.00023 0.08603
12 0.01011 0.01709 0 0.01480 14 0.01244 0.02864 0.00160 0.03689
31 0.05382 0.26950 0 0.00232 26 0.04426 0.16390 0.00219 0.00578
22 0.00030 0.00061 0 0 14 0.00068 0.00160 0.00011 0.00159
1 0.00005 0.00005 0.00005 0 0 — — — 0.00141
31 0.00286 0.00782 0.00094 0.00054 28 0.00334 0.00813 0.00088 0.00135
1 0.00189 0.00189 0.00189 0.00451 3 0.00202 0.00379 0.00069 0.01124
27 0.00074 0.00221 0.00010 0.00133 24 0.00239 0.00551 0.00045 0.00331
12 0.00082 0.00198 0.00010 0.00201 16 0.00155 0.00637 0.00011 0.00501
13 0.00990 0.02482 0.00033 0.02790 10 0.03140 0.06490 0.00239 0.06947
31 0.00346 0.00806 0.00005 0.00220 25 0.00493 0.01395 0.00145 0.00549
td
to
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-10 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT PHOENIX, AZ (in
Phoenix, AZ (040139997)
Parameter
Magnesium (88140)
Manganese (88 132)
Mercury (88142)
Molybdenum (88 134)
Nickel (881 36)
Niobium (88147)
Phosphorous (88 152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88163)
Selenium (88 154)
Silicon (88165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
Sulfur (88 169)
Tantalum (88 170)
Terbium (88 172)
Tin (88160)
Titanium (88161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
Zirconium (88185)
N
20
31
18
18
29
9
0
31
17
1
5
18
31
16
14
29
31
29
9
31
31
17
12
9
31
16
Mean
0.01391
0.00310
0.00085
0.00073
0.00045
0.00069
—
0.10542
0.00051
0.00023
0.00016
0.00042
0.34686
0.00197
0.07840
0.00202
0.39759
0.00822
0.00116
0.00683
0.00879
0.00063
0.00316
0.00041
0.00544
0.00060
URG(5)
Max
0.03612
0.00777
0.00184
0.00212
0.00132
0.00113
—
0.23577
0.00118
0.00023
0.00028
0.00090
1.16593
0.00499
0.17328
0.01026
0.68547
0.01644
0.00334
0.01319
0.02152
0.00184
0.00800
0.00123
0.01262
0.00141
Min
0.00221
0.00085
0.00010
0.00005
0.00005
0.00014
—
0.02623
0.00010
0.00023
0.00010
0.00005
0.07738
0.00014
0.00608
0.00010
0.14079
0.00071
0.00023
0.00221
0.00212
0.00010
0.00090
0.00010
0.00047
0.00010
MDL
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0.00251
0.00137
0.00087
0.00247
0.00097
0.00085
0.00302
0.00420
0.02050
0.00101
0.00265
0.00784
0.00302
0.00717
0.00083
0.00060
0.00554
0.00122
0.00058
0.00144
URG(6)
N Mean Max Min MDL
12 0.01396 0.03523 0.00410 0.00738
31 0.00333 0.00730 0.00099 0.00092
17 0.00077 0.00146 0.00010 0.00175
10 0.00108 0.00311 0.00019 0.00191
26 0.00050 0.00118 0.00005 0.00050
14 0.00066 0.00184 0.00010 0.00168
1 0.00099 0.00099 0.00099 0.00251
31 0.10696 0.28344 0.02926 0.00137
16 0.00041 0.00108 0.00005 0.00087
2 0.00069 0.00128 0.00010 0.00247
1 0.00023 0.00023 0.00023 0.00097
19 0.00041 0.00104 0.00005 0.00085
31 0.34468 1.23567 0.08035 0.00302
19 0.00146 0.00513 0.00023 0.00420
14 0.07773 0.17248 0.00023 0.02050
29 0.00233 0.00923 0.00005 0.00101
31 0.40492 0.71819 0.14642 0.00265
31 0.00699 0.01423 0.00090 0.00784
12 0.00119 0.00416 0.00005 0.00302
31 0.00776 0.01413 0 0.00717
31 0.00860 0.02388 0 0.00083
14 0.00076 0.00146 0 0.00060
18 0.00215 0.00504 0 0.01
12 0.00064 0.00179 0.00005 0
31 0.00658 0.01861 0.00033 0
17 0.00060 0.00127 0.00010 0
N
16
26
19
9
25
15
3
28
13
1
6
7
28
20
13
22
28
27
1
28
28
6
19
8
28
14
Mean
0.02083
0.00435
0.00209
0.00200
0.00370
0.00146
0.00291
0.11410
0.00094
0
0.00057
0.00078
0.47685
0.00502
0.09012
0.00194
0.39844
0.01807
0.00205
0.01739
0.01185
0.00065
0.00576
0.00061
0.00552
0
MetOne
Max
0.05143
0.00887
0.00525
0.00454
0.04763
0.00307
0.00477
0.27725
0.00323
0.00114
0.00103
0.00125
1.41560
0.00958
0.25108
0.01174
0.72924
0.04400
0.00205
0.02781
0.02736
0.00228
0.01920
0.00102
0.01601
0.00324
(7)
Min
0.00139
0.00045
0.00023
0.00034
0.00023
0.00055
0.00089
0.02838
0.00011
0.00114
0.00023
0.00011
0.07708
0.00023
0.01047
0.00011
0.20112
0.00295
0.00205
0.00544
0.00045
0.00011
0.00011
0.00023
0.00023
0.00011
MDL
0.01841
0.00231
0.004
0.00477
0.00125
0.00420
0.00627
0.00341
0.00217
0.00617
0.00243
0.00212
0.00753
0.01048
0.05107
0.00251
0.00662
0.01954
0.00752
0.01787
0.00208
0.00150
0.01380
0.00304
0.001
0.004
td
to
fe
H
6
o
o
H
O
O
H
W
O
O
*The blank spaces mean there are no non-zero, valid measurements.
-------
to
o
o
to
TABLE 3B-11. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT PORTLAND, OR (in
Portland, OR (410510080)
Parameter
PM2 5 Mass (88101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88 109)
Cadmium (88 110)
Calcium (881 11)
Carbonate Carbon (88308)
Cerium (88 11 7)
Cesium (88 11 8)
Chlorine (881 15)
Chromium (881 12)
Cobalt (881 13)
Copper (881 14)
Europium (88121)
Gallium (88124)
Iron (88126)
Lanthanum (88146)
Lead (88128)
N
37
33
35
12
35
35
29
29
29
34
25
33
36
34
22
36
0
17
21
35
27
1
37
1
25
37
21
36
Mean
10.3055
0.34810
0.24866
1.46477
0.78165
1.81885
4.62628
0.69398
5.32026
0.04028
0.0046
0.00161
0.05910
0.00170
0.00252
0.03677
—
0.00733
0.00601
0.13788
0.00126
0.00018
0.01073
0.00094
0.00120
0.06334
0.00934
0.01578
Andersen (5)
Max
61.0285
1.09408
0.71557
16.2389
3.13843
19.2367
11.8141
1.62679
12.4246
0.75854
0.054
0.00967
1.29088
0.00585
0.00466
0.10647
—
0.02011
0.01333
3.28566
0.00445
0.00018
0.26328
0.00094
0.00303
0.24453
0.02025
0.36292
Min
3.14779
0.04244
0.09839
0.06684
0.19594
0.33448
1.94468
0.19218
2.18600
0.00277
0.00032
0
0.00587
0.0002
0.00036
0.00857
—
0.00064
0.00071
0.00087
0.00009
0.00018
0.00050
0.00094
0.00014
0.00726
0.00045
0.00063
MDL
0.04000
0.01500
0.02800
0.01300
0.00800
0.01100
0.13400
0.13400
NA
0.0044
0.00592
0.00099
0.02360
0.00080
0.0042
0.0014
0.13400
0.03450
0.01480
0.00232
0.00063
0.00056
0.00054
0.00451
0.00133
0.00079
0.02790
0.00220
MetOne (6)
N
33
26
31
9
31
31
26
26
26
29
19
26
32
25
18
33
0
18
13
29
24
0
29
2
25
33
23
32
Mean
11.1644
0.43002
0.26273
1.21993
0.86400
1.80200
4.81523
0.71215
5.52738
0.04276
0.00775
0.00256
0.0773
0.00221
0.004
0.04712
—
0.0199
0.0161
0.10324
0.00131
—
0.00999
0.00053
0.00240
0.06408
0.01952
0.01447
Max
39.9458
1.17956
0.56101
9.6779
3.39907
12.7212
10.6439
2.12536
11.4524
0.66708
0.02758
0.00881
0.70015
0.00611
0.01283
0.33552
—
0.0547
0.03794
1.90968
0.007
—
0.16719
0
0.006
0.22954
0.04803
0.22048
Min
4.27350
0.05355
0.11850
0.10714
0.26691
0.40446
2.12891
0.13641
2.26533
0.00083
0.00011
0.00035
0.01070
0.00024
0.00024
0.01378
—
0.00270
0.00107
0.0009
0.0001
—
0.0006
0.0004
0.00011
0.00624
0.00011
0.00035
MDL
0.104
0.017
0.03
0.014
0.008
0.012
0.14600
0.146
NA
0.01088
0.01476
0.00247
0.05876
0.00199
0.01050
0.00347
0.146
0.08603
0.03689
0.00578
0.00159
0.00141
0.00135
0.01124
0.00331
0.00196
0.06947
0.00549
td
to
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-11 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT PORTLAND, OR (in
Portland, OR (410510080)
Parameter
Magnesium (88 140)
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Phosphorous (88152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
Sulfur (88 169)
Tantalum (88 170)
Terbium (88 172)
Tin (88 160)
Titanium (88 161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
N
24
36
21
24
36
19
0
37
16
1
8
23
36
24
36
24
37
33
7
37
36
27
9
14
37
Andersen (5) MetOne (6)
Mean Max Min MDL N Mean Max Min MDL
0.04275 0.86179 0.00073 0.00738 22 0.04338 0.66761 0.00035 0.01841
0.00389 0.01948 0.00041 0.00092 30 0.00432 0.01992 0.00024 0.00231
0.00070 0.00157 0.00004 0.00175 15 0.00264 0.00475 0.00071 0.00437
0.00080 0.00261 0.00004 0.00191 16 0.00216 0.00615 0.00011 0.00477
0.00184 0.00617 0.00022 0.00050 32 0.00931 0.04438 0.00024 0.00125
0.00063 0.00224 0.00004 0.00168 14 0.00205 0.00482 0.00011 0.00420
— — — 0.0025 0 — — — 0.00627
0.42873 13.7123 0.01234 0.00137 33 0.32038 8.76512 0.00603 0.00341
0.00034 0.00071 0.00004 0.00087 10 0.00118 0.00225 0.00012 0.00217
0.00018 0.00018 0.00018 0.00247 1 0.00095 0.00095 0.00095 0.00617
0.00024 0.00073 0.00004 0.00097 6 0.00095 0.00201 0.00035 0.00243
0.00065 0.00234 0.00009 0.00085 19 0.00134 0.00294 0.00012 0.00212
0.07451 0.27605 0.01087 0.00302 32 0.07510 0.25940 0.00978 0.00753
0.00204 0.00462 0.00013 0.00420 20 0.00408 0.01006 0.00024 0.01048
0.15270 0.46990 0.01399 0.02050 32 0.19653 0.51494 0.02382 0.05107
0.01156 0.25846 0.00004 0.00101 13 0.01220 0.14383 0.00011 0.00251
0.61466 6.08678 0.13558 0.00265 33 0.61745 4.09887 0.13664 0.00662
0.00763 0.01849 0.00023 0.00784 31 0.02026 0.04933 0.00188 0.01954
0.00046 0.00141 0.00004 0.00302 2 0.00035 0.00058 0.00011 0.00752
0.00892 0.01603 0.00280 0.00717 33 0.01832 0.02711 0.00613 0.01787
0.00930 0.18710 0.00042 0.00083 33 0.00832 0.11765 0.00118 0.00208
0.00121 0.00462 0.00009 0.00060 16 0.00143 0.00417 0.00012 0.00150
0.00228 0.00478 0.00018 0.00554 15 0.00621 0.01540 0.00105 0.01380
0.00039 0.00081 0.00004 0.00122 10 0.00126 0.00247 0.00011 0.00304
0.01258 0.10924 0.00123 0.00058 32 0.00985 0.05763 0.0001 0.00145
td
to
fe
H
6
o
o
H
O
O
H
W
O
O
*The blank spaces mean there are no non-zero, valid measurements.
-------
to
o
o
to
TABLE 3B-12. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT SALT LAKE CITY, UT (in
Salt Lake City, UT (490353006)
Parameter
PM2 5 Mass (88101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88 109)
Cadmium (88 110)
Calcium (881 11)
Carbonate Carbon (88308)
Cerium (88 11 7)
Cesium (88 11 8)
Chlorine (881 15)
Chromium (881 12)
Cobalt (881 13)
Copper (881 14)
Europium (88121)
Gallium (88124)
Gold (88143)
Lanthanum (88146)
Lead (88128)
Magnesium (88140)
N
33
30
30
5
30
30
22
22
22
33
20
24
32
31
17
33
0
17
12
32
13
2
32
2
29
20
19
33
22
Mean
8.60543
0.41737
0.10337
0.26540
1.07903
0.93003
4.30412
0.69880
5.00292
0.07483
0.00626
0.00210
0.06479
0.00310
0.00546
0.16492
—
0.02896
0.01934
0.01382
0.00105
0.00092
0.00397
0.00093
0.00252
0.00197
0.03008
0.00755
0.01825
Met One (5)
Max
23.3333
2.75328
0.3047
0.42790
8.80071
1.78049
6.37857
1.39183
7.53968
0.39225
0.01536
0.00714
0.16508
0.01029
0.01265
0.58893
—
0.07303
0.04359
0.10220
0.00633
0.00103
0.01427
0.00163
0.00572
0.00550
0.06026
0.04106
0.03693
Min
3.83420
0.07100
0.02387
0.1226
0.18073
0.45655
2.35082
0.24718
2.70476
0.00071
0.00011
0.00011
0.00687
0.00024
0.00011
0.03144
—
0.00105
0.00827
0.00116
0.00011
0.00081
0.00092
0.00024
0.00011
0.00011
0.00129
0.00162
0.00058
MDL
0.10400
0.01700
0.03000
0.01400
0.00800
0.012
0.146
0.146
NA
0.01088
0.01476
0.00247
0.05876
0.00199
0.01050
0.00347
0.14600
0.08603
0.03689
0.00578
0.00159
0.00141
0.00135
0.01124
0.00331
0.00501
0.06947
0.00549
0.01841
Andersen (6)
N
35
31
32
6
32
32
24
24
24
34
22
29
35
34
17
35
0
17
18
35
24
0
35
3
22
8
21
35
27
Mean
8.23244
0.38550
0.09332
0.30784
0.91318
0.87752
4.09455
0.65719
4.75174
0.07018
0.0033
0.0012
0.03573
0.0038
0.00164
0.15305
—
0.01145
0.00719
0.02276
0.00029
—
0.00456
0.00072
0.00080
0.00090
0.01262
0.00654
0.01395
Max
24.7030
2.03933
0.20916
0.59569
7.00222
1.58005
9.02102
1.71189
9.94744
0.30535
0.00780
0.00615
0.12361
0.01245
0.0039
0.40597
—
0.02567
0.01370
0.1633
0.00065
—
0.01507
0.00157
0.00182
0.00207
0.02497
0.06542
0.05318
Min
3.84018
0.04100
0.01739
0.12049
0.13652
0.36993
2.13018
0.16748
2.29766
0.00228
0.00022
0.00013
0.00095
0.00080
0.00004
0.02674
—
0.001
0.0016
0.00193
0
—
0.0016
0
0.00004
0
0.00040
0.00005
0.00088
MDL
0.04
0.01500
0.02800
0.01300
0.00800
0.01100
0.13400
0.13400
NA
0.00436
0.0059
0.00099
0.0236
0.00080
0.00421
0.00139
0.13400
0.03450
0.0148
0.0023
0.001
0.001
0.001
0.0045
0.0013
0.002
0.02790
0.00220
0.00738
td
to
oo
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-12 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS
SALT LAKE CITY, UT (in
AT
Salt Lake City, UT (490353006)
Parameter
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Phosphorous (88152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
Sulfur (88 169)
Tantalum (88 170)
Terbium (88 172)
Tin (88 160)
Titanium (88 161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
N
32
20
12
28
13
5
33
16
1
7
17
33
22
15
19
33
33
7
33
33
2
19
12
33
Met One (5) Andersen (6)
Mean Max Min MDL N Mean Max Min MDL
0.00347 0.00819 0.00046 0.00231 35 0.00328 0.00889 0.00058 0.00092
0.00186 0.00508 0.00011 0.00437 18 0.00070 0.00179 0.00004 0.00175
0.00191 0.00627 0.00011 0.00477 16 0.00087 0.00182 0.00019 0.00191
0.01309 0.17676 0.00011 0.00125 23 0.00038 0.00193 0.00004 0.00050
0.00224 0.00523 0.00023 0.00420 20 0.00068 0.00175 0.00009 0.00168
0.00340 0.00687 0.00011 0.00627 3 0.00309 0.00387 0.00203 0.00251
0.08496 0.44816 0.00721 0.00341 35 0.09978 0.60087 0.00969 0.00137
0.00091 0.00209 0.00011 0.00217 13 0.00052 0.00160 0.00004 0.00087
0.00222 0.00222 0.00222 0.00617 1 0.00058 0.00058 0.00058 0.00247
0.00113 0.00302 0.00011 0.00243 2 0.00020 0.00036 0.00004 0.00097
0.00092 0.00187 0.00011 0.00212 23 0.00042 0.00101 0.00004 0.00085
0.23857 0.95256 0.02825 0.00753 35 0.22324 0.77713 0.03143 0.00302
0.00719 0.02222 0.00046 0.01048 20 0.00197 0.00494 0.00004 0.00420
0.05035 0.18042 0.00035 0.05107 5 0.02972 0.05233 0.01463 0.02050
0.00173 0.00696 0.00024 0.00251 30 0.00151 0.01017 0.00004 0.00101
0.29801 0.64844 0.10711 0.00662 35 0.29797 0.67114 0.08485 0.00265
0.01857 0.03691 0.00024 0.01954 31 0.00568 0.00986 0.00040 0.00784
0.00091 0.00208 0.00024 0.00752 6 0.00066 0.00274 0.00009 0.00302
0.01952 0.03399 0.00210 0.01787 35 0.00818 0.01321 0.00381 0.00717
0.00898 0.02477 0.00223 0.00208 35 0.00742 0.02181 0.00200 0.00083
0.00076 0.00105 0.00046 0.00150 3 0.00028 0.00058 0.00009 0.00060
0.00772 0.01560 0.00011 0.01380 14 0.00217 0.00438 0.00004 0.00554
0.00133 0.00248 0.00024 0.00304 12 0.00055 0.00127 0.00005 0.00122
0.00705 0.02960 0.00070 0.00145 35 0.00827 0.03078 0.0015 0.00058
td
to
VO
fe
H
6
o
o
H
O
O
H
W
O
O
*The blank spaces mean there are no non-zero, valid measurements.
-------
TABLE 3B-13. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT ST. LOUIS, MO (in
to
o
o
to
St. Louis, MO (295100085)
Parameter N
PM2 5 Mass (88101) 35
Ammonium (88301) 38
Sodium Ion (88302) 38
Potassium Ion (88303) 10
Nitrate (88306) 38
Sulfate (88403) 38
Organic Carbon (88305) 24
Elemental Carbon (88307) 24
Total Carbon 24
Aluminum (88 104) 30
Antimony (88102) 17
Arsenic (88103) 28
Barium (88107) 35
Bromine (88 109) 33
Cadmium (88 110) 17
Calcium (881 11) 35
Carbonate Carbon (88308) 0
Cerium (88 11 7) 23
Cesium (881 18) 14
Chlorine (881 15) 16
Chromium (881 12) 23
Cobalt (881 13) 2
Copper (881 14) 34
Europium (88121) 1
Gallium (88124) 28
Mean
15.9460
1.75246
0.13497
0.15767
1.73075
4.23020
4.55208
0.80882
5.36090
0.04142
0.00737
0.00227
0.06549
0.00431
0.00447
0.11230
—
0.02320
0.01361
0.01529
0.00143
0.00023
0.01640
0.00166
0.00208
Met One (5)
Max
36.8015
6.37668
0.36566
0.28898
6.63161
17.5918
6.92448
1.57296
7.87283
0.51741
0.02042
0.00507
0.11098
0.05302
0.01002
0.21063
—
0.05889
0.02719
0.10653
0.00979
0.00035
0.19018
0.00166
0.00425
Andersen (6)
Min MDL N Mean Max Min MDL
5.61856 0.10400 37 15.5672 33.9177 6.51278 0.04000
0.32688 0.01700 40 1.64614 5.48037 0.16619 0.01500
0.01987 0.03000 40 0.13138 0.42148 0.02548 0.02800
0.07387 0.014 15 0.15195 0.36176 0.07339 0.01300
0.25581 0.008 40 1.69734 6.56552 0.24213 0.00800
1.33765 0.012 40 4.15895 16.8408 1.16105 0.01100
2.60132 0.14600 25 4.34840 6.71026 2.54373 0.13400
0.23018 0.14600 25 0.85087 1.74821 0.45032 0.13400
3.16637 NA 25 5.19927 7.86998 3.20344 NA
0.00119 0.01088 26 0.09593 1.81318 0.00285 0.00436
0.00178 0.01476 27 0.00319 0.0074 0.00049 0.00592
0.00035 0.00247 34 0.00150 0.0035 0.00005 0.00099
0.01269 0.05876 36 0.02959 0.05213 0.00066 0.02360
0.00046 0.002 37 0.00426 0.05280 0.001 0.00080
0.00023 0.0105 20 0.00180 0.00792 0 0.00421
0.01564 0.00347 37 0.13481 0.58229 0.04329 0.00139
— 0.146 o — — — 0.13400
0.00143 0.08603 25 0.01048 0.02772 0 0.03450
0.00153 0.03689 20 0.00601 0.01447 0 0.01480
0.00046 0.0058 26 0.02162 0.30948 0 0.00232
0.00012 0.0016 31 0.00136 0.01104 0 0.001
0.00011 0.0014 2 0.00014 0.00023 0.00004 0.001
0.00036 0.0014 37 0.03913 0.71688 0.001 0.001
0.00166 0.01124 2 0.00187 0.00304 0.001 0.00451
0.00011 0.0033 15 0.00071 0.00167 0 0.00133
td
OJ
o
fe
H
6
o
o
H
O
O
H
W
O
O
-------
to
o
o
to
TABLE 3B-13 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT ST. LOUIS, MO (in
St. Louis, MO (295100085)
Parameter
Gold (88 143)
Hafnium (88 127)
Indium (88131)
Indium (88133)
Iron (88 126)
Lanthanum (88 146)
Lead (88 128)
Magnesium (88 140)
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Phosphorous (88152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
N
8
13
14
17
35
17
35
16
34
18
15
34
17
1
35
12
0
4
27
35
24
28
17
Mean
0.00091
0.01301
0.00435
0.00356
0.16336
0.03044
0.01644
0.01989
0.00958
0.00198
0.00160
0.01684
0.00180
0.07769
0.07724
0.00106
—
0.00088
0.00161
0.15945
0.00508
0.08976
0.00115
Met One (5)
Max
0.00199
0.03328
0.01265
0.00684
0.55651
0.07101
0.05871
0.03903
0.05922
0.00389
0.00417
0.10784
0.00341
0.07769
0.30465
0.00178
—
0.00165
0.00389
1.04966
0.01003
0.21981
0.00248
Andersen (6)
Min MDL N Mean Max Min MDL
0.00011 0.00501 1 0.00029 0.00029 0 0.00201
0.00024 0.02605 19 0.00433 0.01215 0 0.0105
0.00059 0.01128 13 0.00178 0.00291 0.00057 0.0045
0.00058 0.00594 10 0.00087 0.00232 0.00014 0.0024
0.02843 0.00196 37 0.18826 1.06552 0.05701 0.00079
0.00472 0.06947 21 0.01341 0.02849 0.00108 0.02790
0.00235 0.00549 37 0.01609 0.08643 0.00227 0.00220
0.00070 0.01841 17 0.01052 0.02089 0.00038 0.00738
0.00118 0.00231 37 0.009 0.05653 0.00097 0.00092
0.00035 0.00437 15 0.001 0.00184 0.00005 0.00175
0.00024 0.00477 19 0.001 0.0021 0.00010 0.00191
0.00189 0.00125 34 0.00135 0.01017 0.00004 0.001
0.00047 0.00420 12 0.00047 0.00100 0.00009 0.00168
0.07769 0.00627 1 0.05253 0.05253 0.05253 0.00251
0.00952 0.00341 37 0.09665 0.70515 0.02424 0.00137
0.00011 0.00217 10 0.00044 0.00092 0.00004 0.00087
— 0.00617 2 0.00111 0.00116 0.00105 0.00247
0.00011 0.00243 3 0.00012 0.00023 0.00004 0.00097
0.00024 0.00212 34 0.00129 0.00458 0.00010 0.00085
0.03410 0.00753 37 0.23356 3.30811 0.03851 0.00302
0.00011 0.01048 24 0.00205 0.00527 0.00018 0.00420
0.00187 0.05107 12 0.06837 0.25709 0.00548 0.02050
0.00011 0.00251 24 0.00101 0.00979 0.00013 0.00101
td
OJ
fe
H
6
o
o
H
O
O
H
W
O
O
-------
TABLE 3 B-13 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT ST. LOUIS, MO (in
to
o
o
to
St. Louis, MO (295100085)
Parameter
Titanium (88 161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
Zirconium (88 185)
N
34
8
12
11
35
15
Mean
0.00639
0.00097
0.00674
0.00104
0.02435
0.00183
Met One (5)
Max
0.02906
0.00295
0.01858
0.00201
0.07338
0.00579
Min
0.00071
0.00011
0.00047
0.00058
0.00107
0.00035
MDL
0.00208
0.00150
0.01380
0.00304
0.00145
0.00359
N
35
13
3
14
37
20
Mean
0.00826
0.00100
0.00094
0.00057
0.04568
0.00088
Andersen
Max
0.09360
0.00403
0.00115
0.00121
0.47977
0.0037
(6)
Min
0.00111
0.00005
0.00057
0.00020
0.00864
0.00004
MDL
0.00083
0.00060
0.00554
0.00122
0.00058
0.0014
td
OJ
to
fe
H
6
o
o
H
O
O
H
W
O
*The blank spaces mean there are no non-zero, valid measurements.
O
-------
to
o
o
to
TABLE 3B-14. SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT SEATTLE, WA (in
Seattle, WA (530330080)
Parameter
PM2 5 Mass (88101)
Ammonium (88301)
Sodium Ion (88302)
Potassium Ion (88303)
Nitrate (88306)
Sulfate (88403)
Organic Carbon (88305)
Elemental Carbon (88307)
Total Carbon
Aluminum (88 104)
Antimony (88102)
Arsenic (88103)
Barium (88 107)
Bromine (88 109)
Cadmium (88 110)
Calcium (881 11)
Carbonate Carbon (88308)
Cerium (88 11 7)
Cesium (88 11 8)
Chlorine (881 15)
Chromium (881 12)
Cobalt (881 13)
Copper (881 14)
Europium (88121)
Gallium (88124)
N
40
32
39
5
39
39
39
39
39
33
28
30
39
36
19
40
0
23
16
34
35
4
38
3
33
Mean
9.4676
0.45565
0.25671
0.55895
0.83427
1.52949
3.82378
0.86143
4.68521
0.02074
0.00646
0.00244
0.05726
0.00196
0.00495
0.04455
—
0.02206
0.01369
0.05698
0.00212
0.00053
0.00350
0.00160
0.00218
Met One (5)
Max
25.1163
1.86358
0.57632
2.31058
3.02535
4.19391
9.31148
2.78426
12.0957
0.17279
0.01577
0.00892
0.21693
0.00560
0.01156
0.11359
—
0.05807
0.04476
0.44694
0.00981
0.00106
0.04612
0.00364
0.00525
Min
3.29897
0.07057
0.06332
0.04885
0.19552
0.54916
1.70716
0.14378
2.13806
0.00024
0.00058
0.00035
0.00094
0.00011
0.00105
0.01447
—
0.00071
0.00011
0.00058
0.00011
0.00011
0.00011
0.00024
0.00024
MDL
0.10400
0.01700
0.03000
0.01400
0.008
0.01200
0.146
0.146
NA
0.0109
0.0148
0.002
0.0588
0.00199
0.01050
0.00347
0.14600
0.08603
0.03689
0.00578
0.00159
0.00141
0.00135
0.01124
0.00331
URG (6)
N
41
41
41
27
41
41
41
41
41
29
21
29
39
41
19
40
0
21
19
39
39
3
41
3
32
Mean
7.58131
0.51602
0.16680
0.14516
0.68270
1.47833
2.50795
0.61197
3.11992
0.0176
0.00311
0.00144
0.03005
0.00199
0.002
0.0337
—
0.01
0.00779
0.03522
0.00174
0.00020
0.00317
0.00077
0.00085
Max
23.7103
1.97357
0.48493
2.33559
2.60029
4.30718
7.09910
1.63835
8.73745
0.20840
0.007
0.004
0.15435
0.00428
0.00551
0.09934
—
0.02387
0.0257
0.30344
0.009
0
0.0449
0.001
0.002
Min
3.04205
0.09762
0.04035
0.01906
0.12613
0.44612
0.93538
0.14849
1.31881
0.00023
0.00028
0.00010
0.00655
0.00033
0.0006
0.00725
—
0.00038
0.00090
0.00005
0.00010
0.00010
0.00010
0.00034
0.00005
MDL
0.04000
0.00700
0.01200
0.00600
0.00300
0.005
0.05900
0.05900
NA
0.00436
0.00592
0.00099
0.02360
0.00080
0.00421
0.00139
0.059
0.0345
0.0148
0.00232
0.0006
0.0006
0.0005
0.00451
0.00133
td
fe
H
6
o
o
H
O
O
H
W
O
O
-------
TABLE 3B-14 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT SEATTLE, WA (in
to
o
o
to
Seattle, WA (530330080)
Met One (5)
Parameter
Gold (88 143)
Hafnium (88 127)
Indium (88131)
Indium (88133)
Iron (88 126)
Lanthanum (88 146)
Lead (88 128)
Magnesium (88 140)
Manganese (88132)
Mercury (88 142)
Molybdenum (88134)
Nickel (88136)
Niobium (88 147)
Phosphorous (88152)
Potassium (88 180)
Rubidium (88 176)
Samarium (88 162)
Scandium (88 163)
Selenium (88 154)
Silicon (88 165)
Silver (88 166)
Sodium (88 184)
Strontium (88 168)
N
20
14
17
21
40
20
39
19
40
22
26
38
16
3
40
13
1
10
22
40
26
37
16
Mean
0.00237
0.00826
0.00475
0.00326
0.07195
0.02539
0.00756
0.02235
0.00471
0.00157
0.00193
0.00667
0.00156
0.00121
0.09503
0.00071
0.00093
0.00090
0.00097
0.06449
0.00550
0.20378
0.00375
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
Max
00647
02125
01196
00941
32351
05843
06970
11371
02406
00459
00515
08607
00470
00200
16303
00200
00093
00247
00199
25778
01152
51094
04213
Min
0.00035
0.00093
0.00047
0.00024
0.01493
0.00520
0.00093
0.00093
0.00035
0.00011
0.00024
0.00071
0.00024
0.00071
0.00745
0.00011
0.00093
0.00011
0.00024
0.01234
0.00035
0.03422
0.00011
MDL
0.00501
0.02605
0.01128
0.00594
0.00196
0.06947
0.00549
0.01841
0.00231
0.00437
0.00477
0.00125
0.00420
0.00627
0.00341
0.00217
0.00617
0.00243
0.00212
0.00753
0.01048
0.05107
0.00251
URG (6)
N
18
17
20
20
41
24
41
28
39
18
28
41
19
0
41
15
1
3
25
40
30
38
25
Mean
0.00097
0.00431
0.00194
0.00117
0.05825
0.01351
0.00662
0.01466
0.00430
0.00108
0.00156
0.00286
0.00068
—
0.09413
0.00037
0.00005
0.00025
0.00044
0.05035
0.00166
0.15366
0.00252
Max
0.003
0.0125
0.005
0.002
0.28782
0.02547
0.07830
0.18727
0.0226
0.00353
0.00513
0.01191
0.00141
—
2.26987
0.00094
0.00005
0.00047
0.00090
0.22823
0.00457
0.39380
0.04359
Min
0.00023
0.00023
0.00005
0.00023
0.01153
0.00523
0.00184
0.00085
0.00014
0.00019
0.0001
0.00038
0.00005
—
0.01046
0.00014
0.00005
0.00010
0.00005
0.00852
0.00005
0.02369
0.00005
MDL
0.00201
0.01050
0.00452
0.00238
0.00079
0.0279
0.00220
0.00738
0.00092
0.00175
0.00191
0.00050
0.00168
0.00251
0.00137
0.00087
0.00247
0.00097
0.00085
0.00302
0.00420
0.02050
0.00101
td
fe
H
6
o
o
H
O
O
H
W
O
O
-------
TABLE 3B-14 (cont'd). SUMMARY STATISTICS FOR THE SPECIATION SAMPLERS AT SEATTLE, WA (in
to
o
o
to
Seattle, WA (530330080)
Met One (5)
Parameter
Titanium (88 161)
Vanadium (88 164)
Wofram(88186)
Yttrium (88183)
Zinc (88167)
Zirconium (88 185)
N
40
25
28
10
40
18
Mean
0.00403
0.00465
0.00603
0.00087
0.00741
0.00146
0
0
0
0
0
0
Max
02453
01665
01537
00152
03048
00423
Min
0.00047
0.00024
0.00035
0.00011
0.00035
0.00011
MDL
0.00208
0.00150
0.01380
0.00304
0.00145
0.00359
URG (6)
N
38
30
19
18
41
23
Mean
0.00321
0.00412
0.00252
0.00072
0.00823
0.00073
Max
0.02604
0.01516
0.00706
0.00250
0.03061
0.00240
Min
0.0007
0.00010
0.00014
0.00005
0.00179
0
MDL
0.00083
0.00060
0.00554
0.00122
0.0006
0.00144
td
fe
H
6
o
o
H
O
O
H
W
O
*The blank spaces mean there are no non-zero, valid measurements.
O
-------
1 REFERENCES
2 Coutant, B.; Stetzer, S. (2001) Evaluation of PM25 speciation sampler performance and related sample collection
3 and stability issues: final report. Research Triangle Park, NC: U.S. Environmental Protection Agency,
4 Office of Air Quality Planning and Standards; report no. EPA-454/R-01-008. Available:
5 http://www.epa.gov/ttn/amtic/pmspec.html [5 April, 2002].
6 Coutant, B.; Zhang, X.; Pivetz, T. (2001) Summary statistics and data displays for the speciation minitrends study:
7 final report. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality
8 Planning and Standards; contract no. 68-D-98-030.
9
April 2002 3B-36 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 3C
2
3 Organic Composition of Particulate Matter
4
5
6 Although organic compounds typically constitute approximately 10 to 70% of the total dry
7 fine particle mass in the atmosphere, organic PM concentrations, composition, and formation
8 mechanisms are poorly understood. This is because particulate organic matter is an aggregate of
9 hundreds of individual compounds spanning a wide range of chemical and thermodynamic
10 properties (Saxena and Hildemann, 1996). The presence of multiphase or "semivolatile"
11 compounds complicates collection of organic particulate matter. Furthermore, no single
12 analytical technique currently is capable of analyzing the entire range of compounds present.
13 Rigorous analytical methods frequently identify only 10 to 20% of the organic mass on the
14 molecular level (Rogge et al., 1993). The data shown in Appendix 3C are meant to complement
15 the data given for the inorganic components of particles in Appendix 6 A of the 1996 PM AQCD
16 (U. S. Environmental Protection Agency, 1996). Table 3C-1 lists a number of recent urban and
17 some rural measurements of particulate organic and elemental carbon in //g of carbon/m3 (//g
18 C/m3). Emphasis is placed on measurements published after 1995. The analysis method and
19 artifact correction procedure, if any, are indicated. Table 3C-2 presents information on recent
20 (post-1990) studies concerning concentrations (in ng C/m3) of particulate organic compounds
21 found at selected U.S. sites.
April 2002 3 C-1 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3C-1. PARTICIPATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in ^ C/m3)
BASED ON STUDIES PUBLISHED AFTER 1995
to
o
o
to
OO
O
to
o
£
H
6
o
o
H
O
O
H
W
0
Reference
URBAN PM2 5
Offenberg and Baker
(2000)
Allen etal. (1999)
Pedersen etal. (1999)
IMPROVE (2000)
Lewtas etal. (2001)
Khwaja (1995)
Christoforou et al.
(2000)
Turpin and Huntzicker
(1995)
Location
Chicago, IL
Uniontown, PA
Boston, MA
Reading, MA (suburban)
Quabbin, MA (rural)
Rochester, NY (urban)
Brockport, NY (rural)
Washington, DC
Seattle, WA
Seattle, WA
Schenectady, NY
Azusa, CA
Long Beach, CA
Central, LA
Rubidoux, LA
San Nicolas, LA
Claremont, CA
Long Beach, CA
Dates
July 1994;
Jan 1995
July-Aug 1990
Jan-Dec 1995
1994-1998
Apr-May 1999
Oct 24-26, 1991
Jan-Dec 1993
Jun-Sept
Nov-Dec 1987
OCMean
(Max)
(0.8-8.4)a
5.8
4.0
2.8
3.3
2.7
3.4
1.8
8.0
9.4
8.9
12.3
9.7
1.6
na (29.4)
na (62.6)
EC Mean TC Mean
(Max) (Max)
2.2 (3.8)
1.7
(0.4-3. 5)a
1.3(3.1)
1.7
0.7
0.5
0.7
0.5
1.1
0.3
1.4
23.2 (49.9)
1.3
1.8
2.7
1.5
1.5
na (9.0)
na (24.6)
Avg.
Time Notes
12 h PM12; Imp; TOT
PMM; Imp; TOT
3 h PM2S; DQQ; TORb
10 min Aeth
24 h PM20; Q; TOT
24 h PM25; QQ; TOR
23 h PM2 5; DQA; EGAC
6 h PMLO; Q; Th
24 h PM2 1, Q; TOR
2h PM25;Q+TQ;TOTd
2-6 h
O
-------
TABLE 3C-1 (cont'd). PARTICULATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in ^g C/m3)
BASED ON STUDIES PUBLISHED AFTER 1995
to
o
o
to
oo
o
fe
H
6
o
o
H
O
O
H
W
O
O
Reference
RURAL PM2 5
Klinedinst and Currie
(1999)
Andrews et al. (2000)
Malm and Gebhart
(1996)
IMPROVE (2000)
Heggetal. (1997)
Cuietal. (1997)
Chow etal. (1996)
Location Dates
Welby, CO Dec 1996-Jan
Brighton, CO 1997
Look Rock, Smoky July-Aug 1995
Mountains, TN
Tahoma Wood, WA June-Aug 1990
Three Sisters Wilderness, OR 1994-1998
Rocky Mountains, CO
Brigantine, NJ
Acadia, MA
Jefferson: James River Face
Wilderness, VA
Glacier, MT
1 50 km East of Mid- Atlantic July 1 996
Coast
(0.02-4 km altitude)
Meadview, AZ Aug 6-15, 1992
Point Reyes, CA July-Aug 1990
Altamont Pass, CA
Pacheco Pass, CA
Crows Landing, CA
Academy, CA
Button- Willow, CA
Edison, CA
Caliente, CA
Sequoia, CA
Yosemite, CA
OCMean
(Max)
5.6(13.4)
3.6 (6.4)
2.2
2.7
1.2
2.6 (7.4)
0.9
1.0
2.0
1.2
3.8
2.4
1.5 (2.7)
4.8 (7.2)
3.2(6.1)
7.4
(12.7)
5.9 (8.7)
6.4
(10.6)
10.0
(12.8)
7.4
(10.7)
5.3 (7.0)
12.1
(25.8)
EC Mean TC Mean Avg.
(Max) (Max) Time Notes
3.3(8.1) 6h PM25;Q;TOR
1.9 (3.6)
0.4 12 h (day) PM2 {; QQ; TOR6
0.1 PM21;Q+TQ;TORd
0.2 PMj 8; Imp; TMO
0.7 (2.2) 12 h PM2 5; QQ; TORf
0.2 24 h PM25;QQ;TOR
0.2
0.5
0.2
0.7
0.4
2.9 (5.4) PMLO; QQ ; EGA6
3.0 12 h PM25; VDQA; EGAC
0.4 (0.6) 5-7
2.6(3.9) hPM25;
1.0 (1.3) Q+TQ;
1.8 (2.5) TORg
1.4 (2.4)
1.9 (2.7)
2.9(4.1)
3.3 (4.4)
1.6 (3.0)
1.9(3.5)
-------
to
o
o
to
TABLE 3C-1 (cont'd). PARTICULATE ORGANIC AND ELEMENTAL CARBON CONCENTRATIONS (in ^g C/m3)
BASED ON STUDIES PUBLISHED AFTER 1995
Reference
Location
Dates
OC Mean
(Max)
EC Mean
(Max)
TCMean
(Max)
Avg.
Time
Notes
oo
O
fe
H
6
o
o
H
O
O
H
W
O
O
RURAL PM2 5
Malm and Day (2000)
PM10
Omar etal. (1999)
Gertleretal. (1995)
Grand Canyon, AZ
Bondville, IL
Bullhead City, AZ
July-Aug 1998
Jan-Dec 1994
Sept 1988-Oct 1989
1.1(1.6)
2.6
6.0 (16.0)
0.10(0.3)
0.2
1.9 (4.0)
24 h PM25; QQ; TORf
24-48 h PM10; Q; TOR
24 h PM10; Q; TOR
Chow etal. (1996)
Lioy and Daisey ( 1987)
Santa Barbara, CA (urban)
Santa Maria, CA (urban)
Santa Ynez, CA (airport)
Gaviota, CA (rural SB)
Watt Road, CA (rural SB)
Anacapa Island, CA
Newark, NJ
Elizabeth, NJ
Camden, NJ
Jan-Dec 1989
1982:
Summer
Winter
Summer
Winter
Summer
Winter
4.1
5.9
2.1
7.1
2.2
5.2
3.0
3.3
1.7
2.3
1.3
2.0
8.8
4.6
3.5
3.4
2.1
3.1
24 h PM10; Q; TOR
PM15; Q
A limited amount of rural data is presented. In some cases, total carbon (TC = OC + EC) is reported. OC concentrations must be multiplied by the average molecular weight per carbon weight to
convert to mass of particulate organic compounds. The location and dates over which sampling occurred are provided. Averaging time refers to the sampling duration. Sampling method: Q - quartz
fiber filter; QQ - two quartz fiber filters in series; Q+TQ - a quartz fiber filter in one port and a Teflon followed by a quartz filter in a parallel port; Imp - cascade impactor; DQQ - denuder followed
by two quartz fiber filters; DQA - denuder followed by quartz fiber filter and adsorbent; VDQA - virtual impactor inlet followed by denuder, quartz filter, and adsorbent. Analysis method is reported
as follows: TOR - thermal optical reflectance; TOT - thermal optical transmittance; TMO - thermal MnO2 oxidation; EGA - evolved gas analysis; Th - Thermal analysis; Aeth - Aethalometer.
na - data not available.
"Range is provided. It should be noted that samples were collected only during elevated pollution episodes and are not representative of average concentrations.
bParticulate OC was considered to be the sum of front and back quartz fiber filters.
°Sum of adsorbent and filter after correction for inlet losses and denuder efficiency.
•"Corrected for adsorption by subtracting the Teflon-quartz back-up filter.
'Reported concentrations are corrected for adsorption by subtracting the quartz (TQ or QQ) back-up filter.
'Sampler contained two quartz fiber filters in series, but publication did not indicate whether the quartz back-up filter was subtracted to correct for adsorption.
8Corrected for adsorption using Micro-Orifice Uniform Deposit Impactor (MOUDI) data from a collocated sampler.
-------
TABLE 3C-2. PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
to
o
o
to
oo
o
fe
H
6
o
o
H
O
O
H
W
O
O
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM2.i
n-Alkanes
n-tricosane
n-tetracosane
n-pentacosane
n-hexacosane
n-heptacosane
n-octacosane
n-nonacosane
n-triacontane
n-hentriacontane
n-dotriacontane
n-tritriacontane
n-tetratriacontane
Total n-alkanes
n-Alkanoic Acids
n-nonanoic acid
n-decanoic acid
n-undecanoic acid
n-dodecanoic acid
n-tridecanoic acid
n-tetradecanoic acid
n-pentadecanoic acid
n-hexadecanoic acid
(palmitic acid)
n-heptadecanoic acid
n-octadecanoic acid
(stearic acid)
n-nonadecanoic acid
n-eicosanoic acid
n-heneicosanoic acid
n-docosanoic acid
n-tricosanoic acid
n-tetracosanoic acid
Los Angeles,
CA
6.7
6.4
11.2
8.2
6.7
3.1
7.1
2.7
12.6
1.5
2.1
0.58
68.9
6.6
2.0
2.8
5.3
4.3
19.7
5.3
140.5
4.7
59.2
1.1
5.1
2.1
8.7
2.0
11.8
Pasadena,
CA
5.4
4.7
9.5
4.3
5.6
2.5
4.7
2.5
9.6
1.5
2.3
0.68
53.3
5.3
2.4
6.0
7.0
4.9
22.2
6.1
127.4
5.2
50.0
1.1
6.1
2.3
9.9
2.5
16.5
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 5 no precut no precut PM[ 9 (urban)
Fresno,
CA
42.3
41.2
29.9
25.0
12.3
33.8
7.39
16.1
2.61
5.02
215.6
0.711
0.905
6.17
9.42
33.7
166
13.6
60.0
10.7
41.2
20.8
160
32.1
205
Bakersfleld,
CA
12.7
14.2
10.7
10.8
5.24
23.6
4.27
9.66
3.50
3.31
98.0
0.164
0.803
1.78
4.01
5.63
54.4
3.77
24.1
2.58
10.4
6.46
43.1
9.71
78.0
Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CO NY Boston, MA CA
19.23 (57.7)
6.04(21.1)
7.77(21.3)
2.08(12.7)
5.62(15.1)
1.26(9.0)
7.70 (20.6)
0.76 (4.6)
5.24(17.9)
0.41 (2.1)
1.49(5.5)
57.9
-------
to
o
o
to
oo
O
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM2.i
n-Alkanoic Acids
(cont'd)
n-pentacosanoic acid
n-hexacosanoic acid
n-heptacosanoic acid
n-octacosanoic acid
n-nonacosanoic acid
n-triacontanoic acid
Total n-alcanoic acids
n-Alkenoic Acids
n-9-hexadecenoic acid
n-9-octadecenoic acid
n-9,12-octadecane-
Los Angeles,
CA
1.3
5.6
0.49
2.7
0.33
1.0
292.6
24.8
Pasadena,
CA
1.6
9.3
0.81
4.9
0.57
2.2
294.3
26.0
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 5 no precut no precut PM[ 9 (urban)
Fresno,
CA
15.4
174
2.56
21.3
1.46
4.32
979.3
18.8
27.1
13.6
Bakersfleld, Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CA CO NY Boston, MA CA
6.59
81.3
2.38
9.65
2.11
5.79
352.7
3.96
3.96
1.83
dienoic acid
Total n-alkenoic acids
24.8
26.0
59.5
9.75
fe
H
6
o
o
H
O
O
H
W
O
n-Alkanals
1-octanal
n-nonanal
n-decanal
n-dodecanal
n-tridecanal
n-tetradecanal
n-pentadecanal
n-hexadecanal
n-heptadecanal
n-octadecanal
Total n-alkanals
3.26
5.7 9.5 19.4 3.01 29.01
23.58
6.01
6.50
9.62
12.47
17.45
24.09
1.84
5.7 9.5 19.4 3.01 133.8
(14.4)
(62.8)
(71.2)
(16.4)
(25.8)
(30.7)
(113.6)
(49.3)
(88.9)
(11.7)
o
-------
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
to
o
o
to
oo
o
fe
H
6
o
o
H
O
O
H
W
O
O
Rogge et al. (1993)a Schauer and Cass (2000)
Jan-Dec 1 982 Dec 26-28, 1995
(annual average) (pollution episode)
PM2, PM2,
Los Angeles, Pasadena, Fresno, Bakersfleld,
CA CA CA CA
n-Alkanols
1-decanol
1-dodecanol
1-tetradecanol
1-pentadecanol
1-hexadecanol
Total n-alkanols
Aliphatic Dicarboxylic
Acids
oxalic acid (C2)
malonic acid 32.7 44.4
(propanedioic)
methylmalonic acid 2.13 nd
(methylpropanedioic)
malonic acid 0.66 1.3
(2-butenedioic)
succinic acid 66.5 51.2
(butanedioic)
methylsuccinic acid 18.0 15.0 24.0 8.80
(methylbutanedioic)
glutaricacid 32.3 28.3 21.3 10.5
(pentanedioic)
methylglutaric acid 19.3 16.6
(methylpentanedioic)
hydroxy butanedioic 14.3 16.0
acid
adipicacid 14.1 14.1 3.39 3.07
(hexanedioic)
pimelicacid 2.22 1.03
(heptanedioic)
suberic acid 3.4 4.1 4.41 13.4
(octanedioic)
axelaicacid 29.0 22.8 19.9 8.22
(nonanedioic)
Total aliphatic 230.3 213.8 77.4 45.0
(ttcarboxvlic acids
Khwaja (1995) Allen et al. (1997)
Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
July24-Aug4, 1989 (semiurban) (urban) Sept 8-9, 1993
no precut no precut PM15 (urban)
Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CO NY Boston, MA CA
8.66(64.1)
21.29(61.7)
13.59(41.4)
4.50(30.1)
27.42(141.1)
75.5
198(360)
84 (107)
102(167)
384
-------
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
0 Rogge et al. (1993)a Schauer and Cass (2000) Khwaja(1995) Allen et al. (1997)
§ Jan-Dec 1982 Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
(annual average) (pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 ! PM2 5 no precut no precut PM[ 9 (urban)
Los Angeles, Pasadena, Fresno, Bakersfleld, Niwot Ridge,
CA CA CA CA CO
Ketocarboxylic Acids
pyruvic acid (C3)
glyoxylic acid (C2)
Total ketocarboxylic
acids
Schenectady, Kenmore Square, Los Angeles Basin,
NY Boston, MA CA
59(103)
44 (68)
103
O
i
oo
fe
H
6
o
o
H
O
O
H
W
O
O
2.3
1.3
3.4
4.8
2.3
4.1
9.97
127
6.68
11.8
0.735
7.95
1.43
2.43
Diterpenoid/Resin Acids
dehydroabietic acid 23.6 22.6 98.5 8.01
abietic acid 30.4 0.784
13-isopropyl-5a- 0.63 1.2
podocarpa-6,8,11,13-
tetraen-16-oic acid
8,15-pimaradien-18-oic 0.44 0.57 0.48 0.03
acid
pimaric acid
isopimaric acid
7-oxodehydroabietic acid
abieta-6,8,11,13,15-
pentaen-18-oic acid
abieta-8,ll,13,15-tetraen- 2.62 0.251
18-oic acid
sandaracopimaric acid 1.6 2.2 8.91 0.525
Total diterpenoid adds 33.3 37.6 296.4 22.15
Aromatic Polycarboxylic
Acids
1,2-benzene-dicarboxylic 60.0 55.7 9.16 6.78
acid (phthalic acid)
1,3-benzene-dicarboxylic 3.4 2.9 3.41 1.98
acid
-------
to
o
o
to
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM,,
Schauer and Cass (2000)
Dec 26-28, 1995
(pollution episode)
Veltkamp et al. (1996)
July24-Aug4, 1989
no precut
Khwaja (1995)
October 1991
(semiurban)
no precut
Allen et al. (1997)
Summer 1994
(urban)
PM,g
Fraser et al. (1998)
Sept 8-9, 1993
(urban)
Los Angeles, Pasadena, Fresno, Bakersfleld, Niwot Ridge, Schenectady,
CA CA CA CA CO NY
Kenmore Square, Los Angeles Basin,
Boston, MA CA
oo
O
fe
H
6
o
o
H
O
O
H
W
O
O
Aromatic Polycarboxylic
Acids (cont'd)
1,4-benzene-dicarboxylic
acid
benzene tricarboxylic
acids
1.5
5.16
14.4
4.48
3.77
4-methy 1-1,2-
benzenedicarboxylic acid
1,2,4-benzene-
tricarboxylic acid
(trimellitic acid)
1,3,5-benzene-
tricarboxylic acid
(trimesic acid)
1,2,4,5-benzene-
tetracarboxylic acid
(pyromellitic acid)
Total aromatic
polycarboxylic acids
Poly cyclic Aromatic
Hydrocarbons
retene
fluoranthene
acephenanthrylene
pyrene
Cr202MWPAH
C2-202 MW PAH
benz [a] anthracene
cyclopenta[c
-------
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
to
o
o
to
p
o
fe
H
6
o
o
H
O
o
H
W
O
O
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM2.i
Los Angeles, Pasadena,
CA CA
Poly cyclic Aromatic
Hydrocarbons (cont'd)
Cr228MWPAH
C2-228 MW PAH
benz [e] acephen-
anthrylene
benzo[fc]fluoranthene 1.15 1.20
benzo[&]fluoranthene 1.23 0.85
benzo [/] fluoranthene
benzo[e]pyrene 0.97 0.93
benzo[a]pyrene 0.42 0.44
perylene
methyl-substituted 252
MWPAH
mdeno[l,2,3-cd]-pyrene 0.37 0.42
mdeno[l,2,3-cd]- 1.05 1.09
fluoranthene
benzo[g/n']perylene 4.47 4.43
anthanthrene
coronene
Total poly cyclic aromatic 11.66 11.10
hydrocarbons
Oxygenated PAHs/
Poly cyclic Aromatic
Ketones/Quinones
1 ,4-naphthoquinone
1 -acenaphthenone
9-fluorenone
1,8-naphthalic anhydride
phenanthrenequinone
phenalen-9-one
anthracene-9, 1 0-dione
methy lanthracene- 9,10-
dione
1 lH-benzo[a]fluoren-l 1-
one
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 5 no precut no precut PM[ 9 (urban)
Fresno,
CA
17.6
8.69
10.7
3.62
7.20
8.23
1.50
6.84
1.36
9.75
0.180
139.57
Bakersfleld, Niwot Ridge,
CA CO
5.35
2.13
2.48
0.499
1.98
1.77
0.246
2.56
0.764
3.49
0.131
34.40
Schenectady, Kenmore Square, Los Angeles Basin,
NY Boston, MA CA
0.34(2.16)
0.09 (0.46)
0.20(1.00)
0.22(1.07)
0.02(0.10)
0.22(1.00)
0.14(0.80)
0.05(0.51)
0.10(0.88)
0.29(1.38)
0.10(0.46)
0.77(4.23)
3.77
0.26
2.07 0.29(1.04)
1.77 0.41(1.65)
0.43
0.53(2.23)
0.36(1.14)
0.09 (0.24)
1.03
-------
to
o
o
to
O
fe
H
6
o
o
H
O
O
H
W
O
O
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
PM2.i
Los Angeles, Pasadena,
CA CA
Schauer and Cass (2000)
Dec 26-28, 1995
(pollution episode)
PM2.5
Fresno, Bakersfleld,
CA CA
Veltkamp et al. (1996)
July24-Aug4, 1989
no precut
Niwot Ridge,
CO
Khwaja (1995)
October 1991
(semiurban)
no precut
Schenectady,
NY
Allen et al. (1997)
Summer 1994
(urban)
PML9
Kenmore Square,
Boston, MA
Fraser et al. (1998)
Sept 8-9, 1993
(urban)
Los Angeles Basin,
CA
Oxygenated PAHs/
Poly cyclic Aromatic
Ketones/Quinones
(cont'd)
7H-benzo [c] fluoren-7-one
1 lH-benzo[6]fluoren-l 1-
one
1 H-phenalen-1 -one
benzanthrone
5,12-naphthacene-quinone
7H-benz[rfe]-anthracen-7- 0.81
one
benz [de] anthracene-7-
dione
benz[a]anthracene-7,12- 0.21
dione
cyclopenta[rfe/|phen-
anthrone
benzo[«/]pyren-6-one 0.80
6H-benzo[crf]pyrene-6-
one
benzo [a]pyrene-6,12-
dione
Total poly cyclic aromatic 1.82
ketones/quinones
7.96
0.588
0.37
0.85
1.18
0.32
0.84
0.25
1.24
2.33
15.76
2.07
1.34
0.096
9.72
0.20(1.00)
0.09(0.31)
0.05(0.14)
0.54 (2.47)
2.56
Steroids
cholesterol nd
Substituted Phenols
/>-benzenediol
OT-benzenediol
hydroxybenzaldehydes
Total substituted phenols
1.9
3.46
7.59
2.64
13.69
nd
nd
0.604
0.604
-------
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
to
o
o
to
p
to
fe
H
6
o
o
H
O
O
H
W
O
O
Rogge et al. (1993)a
Jan-Dec 1982
(annual average)
Los Angeles, Pasadena,
CA CA
Guaiacol and Substituted
Guaiacols
guaiacol
4-methylguaiacol
trans-isoeugenol
vanillin
acetovanillone
guaiacyl acetone
coniferyl aldehyde
Total guaiacol and
substituted guaiacols
Syringol and Substituted
Syringols
syringol
4-methylsyringol
4-ethylsyringol
4-propylsyringol
4-propenylsyringol
syringaldehyde
acetosyringone
acetonylsyringol
propionylsyringol
butyrylsyringol
sinapyl aldehyde
Total syringol and
substituted syringols
Sugars
levoglucosan
other sugars
Total sugars
Schauer and Cass (2000) Khwaja (1995) Allen et al. (1997)
Dec 26-28, 1995 Veltkamp et al. (1996) October 1991 Summer 1994 Fraser et al. (1998)
(pollution episode) July 24-Aug 4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 5 no precut no precut PM[ 9 (urban)
Fresno,
CA
0.889
0.606
1.45
26.8
3.23
10.8
47.0
90.78
1.16
1.72
2.28
0.871
4.38
135
171
406
32.1
15.3
15.9
785.7
7590
1070
8660
Bakersfleld, Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CA CO NY Boston, MA CA
0.832
0.387
1.04
6.05
0.705
4.29
nd
13.30
0.845
1.77
2.39
nd
1.40
44.5
55.7
68.1
16.2
6.18
197.1
1100
171
1271
-------
TABLE 3C-2 (cont'd). PARTICULATE ORGANIC COMPOUND CONCENTRATIONS (in ng C/m3) BASED ON STUDIES
PUBLISHED AFTER 1990 AT SELECTED SITES
o
o
to
Other Compounds
divanillyl
divanillyl methane
vanillylmethylguaiacol
Total other
Rogge et al. (1993)a Schauer and Cass (2000) Veltkamp et al. Khwaja (1995) Allen et al. (1997)
Jan-Dec 1982 Dec 26-28, 1995 (1996) October 1991 Summer 1994 Fraser et al. (1998)
(annual average) (pollution episode) July 24-Aug4, 1989 (semiurban) (urban) Sept 8-9, 1993
PM2 ! PM2 5 no precut no precut PM[ 9 (urban)
Los Angeles, Pasadena, Fresno,
CA CA CA
19.4
2.39
3.24
25.0
Bakersfield, Niwot Ridge, Schenectady, Kenmore Square, Los Angeles Basin,
CA CO NY Boston, MA CA
3.18
nd
0.568
3.75
O
fe
H
6
o
o
H
O
O
H
W
O
O
N-Containing Compounds
3 -methoxypyridine
isoquinoline
1 -methoxypyridine
1 ,2-dimethoxy-4-nitro-
benzene
dihydroxynitrobenzene
Total N-containing
compounds
Total Quantified Organic
Compound Mass
Total Organic Compound
Mass
Percent of Organic Mass
Quantified
Percent of Organic Mass
Extractable and Elutable
0.86
1.1
0.27
1.8
4.03
789
1.4
1.1
0.24
3.9
6.64
764
8-15% (a) 8-15% (a)
45-60% (a) 45-60% (a)
11410
55700
20%
30%
2075
18700
11%
21%
1.62(10.52)
1.62
267
487
<3%
10
Mean values are provided with maximum concentrations in parentheses.
"Rogge et al. (1993) summarized these percentages for all four Los Angeles Basin sampling sites (West LA, Downtown LA, Pasadena, and Rubidoux). Only Downtown LA and Pasadena data are
shown here.
-------
1 REFERENCES
2 Allen, J. O.; Dookeran, N. M; Taghizadeh, K.; Lafleur, A. L.; Smith, K. A.; Sarofim, A. F. (1997) Measurement of
3 oxygenated polycyclic aromatic hydrocarbons associated with a size-segregated urban aerosol. Environ. Sci.
4 Technol. 31:2064-2070.
5 Allen, G. A.; Lawrence, J.; Koutrakis, P. (1999) Field validation of a semi-continuous method for aerosol black
6 carbon (aethalometer) and temporal patterns of summertime hourly black carbon measurements in
7 southwestern PA. Atmos. Environ. 33: 817-823.
8 Andrews, E.; Saxena, P.; Musarra, S.; Hildemann, L .M.; Koutrakis, P.; McMurry, P. H.; Olmez, L; White, W. H.
9 (2000) Concentration and composition of atmospheric aerosols from the 1995 SEAVS experiment and a
10 review of the closure between chemical and gravimetric measurements. J. Air Waste Manage. Assoc.
11 50: 648-664.
12 Chow, J. C.; Watson, J. G.; Lu, Z.; Lowenthal, D. H.; Frazier, C. A.; Solomon, P. A.; Thuillier, R. H.; Magliano, K.
13 (1996) Descriptive analysis of PM2 5 and PM10 at regionally representative locations during
14 SJVAQS/AUSPEX. In: Parrish, D.; Trainer, M.; Rao, S. T.; Solomon, P. A., eds. A&WMA international
15 specialty conference on regional photochemical measurements and modeling, part 2; November 1993;
16 SanDiego, CA. Atmos. Environ. 30: 2079-2112.
17 Christoforou, C. S.; Salmon, L. G.; Hannigan, M. P.; Solomon, P. A.; Cass, G. R. (2000) Trends in fine particle
18 concentration and chemical composition in southern California. J. Air Waste Manage. Assoc. 50: 43-53.
19 Cui, W.; Machir, J.; Lewis , L.; Eatough, D. J.; Eatough, N. L. (1997) Fine paniculate organic material at Meadview
20 during the project MOHAVE summer intensive study. J. Air Waste Manage. Assoc. 47: 357-369.
21 Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T.; Rasmussen, R. A. (1998) Air quality model evaluation data for
22 organics. 5. C6-C22 nonpolar and semipolar aromatic compounds. Environ. Sci. Technol. 32: 1760-1770.
23 Gertler, A. W.; Lowenthal, D. A.; Coulombe, W. G. (1995) PM10 source apportionment study in Bullhead City,
24 Arizona. J. Air Waste Manage. Assoc. 45: 75-82.
25 Hegg, D. A.; Livingston, J.; Hobbs, P. V.; Novakov, T.; Russell, P. (1997) Chemical apportionment of aerosol
26 column optical depth off the mid-Atlantic coast of the United States. J. Geophys. Res. 102: 25,293-25,303.
27 IMPROVE: interagency monitoring of protected visual environments [database]. (2000) [Data on paniculate
28 organic and elemental carbon concentrations after 1995]. Fort Collins, CO: National Park Service (NFS);
29 Cooperative Institute for Research in the Atmosphere (CIRA). Available at:
30 http://vista.cira.colostate.edu/improve/ [2001, January 26].
31 Khwaja, H. (1995) Atmospheric concentrations of carboxylic acids and related compounds at a semiurban site.
32 Atmos. Environ. 29: 127-139.
3 3 Klinedinst, D. B.; Currie, L. A. (1999) Direct quantification of PM2 5 fossil and biomass carbon within the northern
34 front range air quality study's domain. Environ. Sci. Technol. 33: 4146-4154.
35 Lewtas, J.; Pang, Y.; Booth, D.; Reimer, S.; Eatough, D. J.; Gundel, L. A. (2001) Comparison of sampling methods
36 for semi-volatile organic carbon associated with PM2 5. Aerosol. Sci. Technol. 34: 9-22.
37 Lioy, P. J.; Daisey, J. M. (1987) Toxic air pollution: a comprehensive study of non-criteria air pollutants. Chelsea,
38 MI: Lewis Publishers.
39 Malm, W. C.; Day, D. E. (2000) Optical properties of aerosols at Grand Canyon National Park. Atmos. Environ.
40 34:3373-3391.
41 Malm, W. C.; Gebhart, K. A. (1996) Source apportionment of organic and light-absorbing carbon using receptor
42 modeling techniques. Atmos. Environ. 30: 843-855.
43 Offenberg, J. H.; Baker, J. E. (2000) Aerosol size distributions of elemental and organic carbon in urban and
44 over-water atmospheres. Atmos. Environ. 34: 1509-1517.
45 Omar, A. H.; Biegalski, S.; Larson, S. M.; Landsberger, S. (1999) Paniculate contributions to light extinction and
46 local forcing at a rural Illinois site. Atmos. Environ. 33: 2637-2646.
47 Pedersen, D. U.; Durant, J. L.; Penman, B. W.; Crespi, C. L.; Hemond, H. F.; Lafleur, A. L.; Cass, G. R. (1999)
48 Seasonal and spatial variations in human cell mutagenicity of respirable airborne particles in the northeastern
49 United States. Environ. Sci. Technol. 33: 4407-4415.
50 Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.; Simoneit, B. R. T. (1993) Quantification of urban
51 organic aerosols at a molecular level: identification, abundance and seasonal variation. Atmos. Environ.
52 Part A 27: 1309-1330.
53 Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
54 literature and applications of thermodynamics to identify candidate compounds. J. Atmos. Chem. 24: 57-109.
April 2002 3 C-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 Schauer, J. J.; Cass, G. R. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants
2 using organic compounds as tracers. Environ. Sci. Technol. 34: 1821-1832.
3 Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
4 primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
5 U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
6 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
7 Veltkamp, P. R.; Hansen, K. J.; Barkley, R. M; Sievers, R. E. (1996) Principal component analysis of summertime
8 organic aerosols at Niwot Ridge, Colorado. J. Geophys. Res. [Atmos.] 101: 19,495-19,504.
9
April 2002 3 C-15 DRAFT-DO NOT QUOTE OR CITE
-------
i APPENDIX 3D
2
3 Composition of Particulate Matter Source Emissions
4
5
6 This appendix includes discussions of the elemental composition of emissions from various
7 source categories discussed in Table 3-8. Discussions in this appendix incorporate material
8 dealing with the inorganic components of source emissions from Chapter 5 of the 1996 PM
9 AQCD (U. S. Environmental Protection Agency, 1996), updates to that material, and material
10 describing the composition of organic components in source emissions. Primary emphasis is
11 placed in the discussions on the composition of PM25 sources.
12
13 Soil and Fugitive Dust
14 The compositions of soils and average crustal material are shown in Table 3D-1 (adapted
15 from Warneck, 1988). Two entries are shown as representations of average crustal material.
16 Differences from the mean soil composition shown can result from local geology and climate.
17 Major elements in both soil and crustal profiles are Si, Al, and Fe, which are found in the form of
18 various minerals. In addition, organic matter constitutes a few percent, on average, of soils.
19 In general, the soil profile is similar to the crustal profiles, except for the depletion of soluble
20 elements such as Ca, Mg, Na, and K. It should be noted that the composition of soils from
21 specific locations can vary considerably from these global averages, especially for elements like
22 Ca, Mg, Na, and K.
23 Fugitive dust emissions arise from paved and unpaved roads, building construction and
24 demolition, parking lots, mining operations, storage piles, feed lots, grain handling, and
25 agricultural tilling, in addition to wind erosion. Figure 3D-1 shows examples of size
26 distributions in dust from paved and unpaved roads, agricultural soil, sand and gravel, and
27 alkaline lake bed sediments, which were measured in a laboratory resuspension chamber as part
28 of a study in California (Chow et al., 1994). This figure shows substantial variation in particle
29 size among some of these fugitive dust sources. The PMX 0 abundance (6.9%) in the total
30 suspended PM (TSP) from alkaline lake bed dust is twice its abundance in paved and unpaved
31 road dust. Approximately 10% of the TSP is in the PM2 5 fraction and approximately 50% of TSP
32 is in the PM10 fraction. The sand/gravel dust sample shows that 65% of the mass is in particles
April 2002 3D-1 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3D-1. AVERAGE ABUNDANCES OF MAJOR ELEMENTS IN
SOIL AND CRUSTAL ROCK
Element
Si
Al
Fe
Ca
Mg
Na
K
Ti
Mn
Cr
V
Co
Elemental
Soil
(1)
330,000
71,300
38,000
13,700
6,300
6,300
13,600
4,600
850
200
100
8
Abundances (ppmw)
(2)
277,200
81,300
50,000
36,300
20,900
28,300
25,900
4,400
950
100
135
25
Crustal Rock
(3)
311,000
77,400
34,300
25,700
33,000
31,900
29,500
4,400
670
48
98
12
Source: (1) Vinogradov (1959); (2) Mason (1966); (3) Turekian (1971), Model A; as quoted in Warneck (1988).
1 larger than the PM10 fraction. The PM2 5 fraction of TSP is approximately 30 to 40% higher in
2 alkaline lake beds and sand/gravel than in the other soil types. The tests were performed after
3 sieving and with a short (<1 min) waiting period prior to sampling. It is expected that the
4 fraction of PMj 0 and PM2 5 would increase with distance from a fugitive dust emitter as the larger
5 particles deposit to the surface faster than do the smaller particles.
6 The size distribution of samples of paved road dust obtained from a source characterization
7 study in California is shown in Figure 3D-2. As might be expected, most of the emissions are in
8 the coarse size mode. The chemical composition of paved road dust obtained in Denver, CO,
9 during the winter of 1987-1988 is shown in Figure 3D-3. The chemical composition of paved
10 road dust consists of a complex mixture of particulate matter from a wide variety of sources.
11 Hopke et al. (1980) found that the inorganic composition of urban roadway dust in samples from
April 2002 3D-2 DRAFT-DO NOT QUOTE OR CITE
-------
100
Paved Unpaved Agricultural Soil/Gravel Alkaline
Road Dust Road Dust Soil Lake Bed
CZ]<1.0|jm CZI<2.5|jm
TSP
Figure 3D-1. Size distribution of particles generated in a laboratory resuspension
chamber.
Source: Chow etal. (1994).
1 Urbana, IL, could be described in terms of contributions from natural soil, automobile exhaust,
2 rust, tire wear, and salt. Automobile contributions arose from exhaust emissions enriched in Pb;
3 from rust as Fe; tire wear particles enriched in Zn; brake linings enriched in Cr, Ba, and Mn; and
4 cement particles derived from roadways by abrasion. In addition to organic compounds from
5 combustion and secondary sources, road dust also contains biological material such as pollen and
6 fungal spores.
Very limited data exist for characterizing the composition in organic compounds found in
resuspended paved road dust and soil dust. The only reported measurements are from Rogge
et al. (1993a) and Schauer and Cass (2000), which consist of data for the fine particle fraction.
The resuspended road dust sample analyzed by Rogge et al. (1993a) was collected in Pasadena,
CA, during May of 1988. The sample analyzed by Schauer and Cass (2000) is a composite
April 2002
3D-3
DRAFT-DO NOT QUOTE OR CITE
-------
99.2%
100 -i
Q_
CO
I
0)
D_
80-
60-
40-
20-
52.3%
\(<10u)
10.7%
(<2.5u)
4.5%
92.8%
82.7%
(<2.5u)
81.6%
95.8%
93.1%
(<2.5u)
92.4%
34.9%
Road and Agricultural Residential Diesel
Soil Dust Burning Wood Truck
Combustion Exhaust
Crude Oil Construction
Combustion Dust
Code:
>10u
-10|J
1u-2.<
IM
Figure 3D-2. Size distribution of California source emissions, 1986.
Source: Houcketal. (1989, 1990).
April 2002 3D-4 DRAFT-DO NOT QUOTE OR CITE
-------
10'
10
10'
.£>.
'^^&^^«$f<&<$i&tff4f&fS
" O O V0y C^ *s?^ Cj X^1* "C* ^ j*S^> ^^Cj t>^ £& sifc^ ».
-------
TABLE 3D-2. SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
PRESENT IN FINE PARTICLE ROAD DUST SAMPLE
Source
Pasadena Road Dust
(Rogge et al., 1993a)
San Joaquin Valley
Road Dust (Schauer
and Cass, 2000)
Contribution to Dominant Contributors to
Compound Class Paniculate Mass (%) Emissions of Compound Class
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Petroleum biomarkers
PAH
n-Alkanals
n-Alkanols
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
0.13
0.37
0.028
0.017
0.0059
0.046
0.021
0.023
0.23
0.095
CIY, C19, C21
Palmitic acid and stearic acid
Oleic acid and linoleic Acid
Hopanes and steranes
No dominant compounds
Octacosanol and triacontanal
Hexacosanol and octacosanol
No dominant compounds
Palmitic acid and stearic acid
Oleic acid, linoleic acid, and
hexadecenoic acid
1 Stationary Sources
2 The elemental composition of primary particulate matter emitted in the fine fraction from a
3 variety of power plants and industries in the Philadelphia area is shown in Table 3D-3 as a
4 representative example of emissions from stationary fossil combustion sources (Olmez et al.,
5 1988). Entries for the coal fired power plant show that Si and Al, followed by sulfate, are the
6 major primary constituents produced by coal combustion; whereas fractional abundances of
7 elemental carbon were much lower and organic carbon species were not detected. Sulfate is the
8 major parti culate constituent released by the oil fired power plants examined in this study; and,
9 again, elemental and organic carbon are not among the major species emitted. Olmez et al.
10 (1988) also compared their results to a number of similar studies and concluded that their data
11 could have much wider applicability to receptor model studies in other areas with some of the
12 same source types. The high temperature of combustion in power plants results in the almost
13 complete oxidation of the carbon in the fuel to CO2 and very small amounts of CO. Combustion
14 conditions in smaller boilers and furnaces allow the emission of unburned carbon and sulfur in
April 2002
3D-6
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3D-3. COMPOSITION OF FINE PARTICLES RELEASED BY VARIOUS STATIONARY
SOURCES IN THE PHILADELPHIA AREA
to
o
o
to
OJ
O
o
l>
H
6
o
o
H
0
0
H
W
0
/^•\
Species Eddystone Coal-
(Units) Fired Power Plant
C-v (%)
C-e (%)
NH4 (%)
Na (%)
Al (%)
Si (%)
P (%)
S (%)
c/~\ /o/ \
O\-/4 ^ /oj
Cl (%)
K(%)
Ca (%)
Sc (ppm)
Ti (%)
V (ppm)
Cr (ppm)
Mn (ppm)
Fe (%)
Co (ppm)
Ni (ppm)
ND
0.89 ±0.12
1.89 ±0.19
0.31 ±0.03
14 ±2
21.8±1.6
0.62 ±0.10
3.4 ±0.6
11.9±1.2
0.022 ±0.11
1.20 ±0.09
1.4 ±0.5
42 ±2
1.1 ±0.2
550 ± 170
390 ± 120
290 ± 15
7.6 ±0.4
93 ±10
380 ±50
Oil-Fired Power Plants
; N
3
3
3
3
9
9
9
3
3
9
3
3
3
3
3
3
3
3
9
Eddystone
2.7 ±1.2
7.7 ±1.5
3.5 ±1.6
3.0 ±0.8
0.45 ±0.09
1.9 ±0.6
1.5 ±0.4
11±2
40 ±4
0.019 ±0.009
0.16 ±0.05
3.6 ±1.0
0.17 ±0.02
0.040 ± 0.044
11500 ±3000
235 ±10
380 ± 40
1.6 ±0.2
790 ± 150
15000 ± 5000
N
3
3
3
3
3
9
9
9
3
2
9
3
3
9
3
3
3
3
3
9
Schuylkill
0.75 ±0.63
0.22 ±0.17
3.7 ± 1.7
3. 3 ±0.8
0.94 ±0.08
2.6 ±0.4
1.0 ±0.2
13 ±1
45 ±7
ND
0.21 ±0.03
2.3 ±1.0
0.47 ±0.02
0.12 ±0.02
20,000 ± 3000
230 ± 70
210 ±50
1.7 ±0.4
1100 ±200
19000 ± 2000
N
4
4
4
3
3
11
11
11
4
11
3
3
11
3
3
3
3
3
11
Secondary
Al Plant
1.6 ± 1.5
0.18±0.10
2.2 ±0.9
16.3 ±0.8
1.74 ±0.09
3.1 ±2.2
0.45 ± 0.27
3±4
5.9 ±2
21±4
10.9 ±1.5
0.12 ±0.09
0.092 ±0.039
0.024 ± 0.003
36 ±7
410 ±20
120 ±15
0.31 ±0.02
13 ±2
300 ± 100
N
2
2
2
1
1
2
2
2
2
1
2
2
1
2
1
1
1
1
1
2
Fluid Cat.
Cracker
ND
0.16 ±0.05
0.43 ± 0.22
0.38 ±0.05
6.8 ± 1.2
9.8 ±20.0
ND
4.2 ± 12.6
38 ±4
ND
0.031 ±0.005
0.030 ±0.004
2.7 ±0.4
0.38 ±0.1
250 ± 70
59 ±8
14 ±3
0.20 ±0.03
15 ±2
220 ± 30
N
3
3
3
3
9
9
3
9
9
3
3
3
3
3
9
3
9
Municipal
Incinerator
0.57 ± 0.26
3.5 ±0.2
0.36 ±0.07
6.6 ±3. 5
0.25 ±0.10
1.7 ±0.3
0.63 ±0.12
2.9 ±0.8
6.8 ±2.3
29 ±5
7.6 ±2.3
0.23 ±0.10
0.11 ±0.02
0.030 ±0.015
8.6 ±5. 3
99 ±31
165 ± 40
0.22 ±0.05
3.7 ±0.8
290 ± 40
N
4
4
4
3
3
10
10
10
4
3
10
10
1
10
2
3
3
3
3
10
-------
TABLE 3D-3 (cont'd). COMPOSITION OF FINE PARTICLES RELEASED BY VARIOUS STATIONARY
SOURCES IN THE PHILADELPHIA AREA
to
o
o
to
OJ
O
oo
o
§
H
O
o
H
0
0
H
W
0
O
Species
(units)
Cu (ppm)
Zn (%)
As (ppm)
Se (ppm)
Br (ppm)
Rb (ppm)
Sr (ppm)
Zr (ppm)
Mo (ppm)
Ag (ppm)
Cd (ppm)
In (ppm)
Sn (ppm)
Sb (ppm)
Cs (ppm)
Ba (ppm)
La (ppm)
Ce (ppm)
Nd (ppm)
Sm (ppm)
Eddy stone
Coal-Fired
Power Plant
290 ± 20
0.041 ±0.005
640 ± 80
250 ± 20
35 ±8
190 ± 80
1290 ± 60
490 ± 190
170 ± 60
ND
ND
0.71 ±0.04
ND
a
9.2 ±0.9
ND
120 ± 10
180 ±10
80 ±26
23 ±2
Oil-Fired Power Plants
N
9
3
3
3
3
1
9
9
2
2
2
3
2
3
3
Eddystone
980 ± 320
1.3 ±0.3
33 ±6
26 ±9
90 ±60
ND
160 ± 50
140 ± 180
930 ±210
ND
ND
ND
320 ±230
370 ±410
ND
1960 ± 100
130 ±30
89 ±23
28 ±5
3.7 ±0.7
N
9
3
1
3
9
9
9
3
9
3
3
3
3
2
3
Schuylkill
1100 ±500
0.78 ±0.30
50 ±16
23 ±7
45 ±17
ND
280 ± 70
100 ± 120
1500 ±300
ND
ND
ND
200 ± 80
1020 ± 90
ND
2000 ± 500
450 ±30
360 ± 20
230 ± 20
20.5 ± 1.5
N
11
3
3
3
11
11
11
3
11
3
3
3
3
3
3
Secondary
Al Plant
450 ± 200
0.079 ± 0.006
15 ±6
66 ±3
630 ± 70
97 ±38
ND
ND
ND
ND
ND
ND
550 ± 540
6100 ±300
ND
ND
19 ±2
ND
ND
ND
N Fluid Cat. Cracker
2 14 ±8
1 0.0026 ± 0.0007
1 ND
1 15± 1
2 5.6 ±1.8
1 ND
36 ±6
130 ±50
ND
ND
ND
ND
2 ND
1 7.7 ±1.5
ND
290 ± 90
1 3300 ±500
2700 ± 400
1800 ± 250
170 ± 20
N
9
3
3
9
9
2
3
2
3
3
3
3
Municipal
Incinerator
1300 ±500
10.4 ±0.5
64 ±34
42 ± 16
2300 ± 800
230 ±50
87 ±14
ND
240 ± 130
71 ±15
1200 ± 700
4.9 ±1.4
6700 ± 1900
1300 ± 1000
5.9 ±3.0
ND
1.1±0.5
ND
ND
ND
N
3
3
3
3
10
2
10
10
3
3
3
10
3
3
1
-------
TABLE 3D-3 (cont'd). COMPOSITION OF FINE PARTICLES RELEASED BY VARIOUS
STATIONARY SOURCES IN THE PHILADELPHIA AREA
o
o
10 Species
(units)
Eu (ppm)
Gd (ppm)
Tb (ppm)
Yb (ppm)
Lu (ppm)
Hf(ppm)
Ta (ppm)
W (ppm)
Au (ppm)
w Pb (%)
O
Th (ppm)
% mass
Eddystone Coal-
Fired Power
Plant
5.1 ±0.5
ND
3. 3 ±0.3
10.3 ±0.5
ND
5.8 ±0.8
ND
20 ±8
ND
0.041 ±0.004
24 ±2
24 ±2
Oil-Fired Power Plants
N
3
3
1
3
1
9
3
6
Eddystone N
ND
ND
ND
ND
ND
0.39 ±0.07 1
ND
60 ±5 2
0.054 ±0.017 2
1.8 ±0.6 9
1.9 ±0.5 2
93. 5 ±2.5 6
Schuylkill
0.65 ±0.23
ND
0.90 ±0.29
ND
ND
ND
ND
ND
ND
1.0 ±0.2
ND
96 ±2
Secondary
N Al Plant
3 ND
ND
3 ND
ND
ND
ND
ND
ND
ND
11 0.081 ±0.014
ND
6 81 ±10
Fluid Cat.
N Cracker
4.9 ±0.7
71 ±10
8.9 ±1.3
3.7 ±0.4
0.59 ±0.17
0.99 ±0.08
0.56 ±0.10
ND
ND
2 0.0091 ±0.0021
6.2 ±0.7
2 97 ±2
N
3
3
3
3
3
3
3
9
3
7
Municipal
Incinerator
ND
ND
ND
ND
ND
ND
ND
ND
0.56 ±0.27
5.8 ±1.2
ND
89 ±2
N
3
10
7
o
H
O
c
o
H
W
O
&
O
aOmitted because of sample contamination.
N = Number of samples.
ND = Not detected.
The "% mass" entries give the average percentage of the total emitted mass found in the fine fraction.
Source: Adapted from Olmez et al. (1988).
-------
1 more reduced forms such as thiophenes and inorganic sulfides. A number of trace elements are
2 greatly enriched over crustal abundances in different fuels, such as Se in coal and V, Zn, and Ni
3 in oil. In fact, the higher V content of the fuel oil than in coal could help account for the higher
4 sulfate seen in the profiles from the oil-fired power plant compared to the coal-fired power plant
5 because V at combustion temperatures found in power plants is known to catalyze the oxidation
6 of reduced sulfur species. During combustion at lower temperatures, the emission of reduced
7 sulfur species also occurs. For example, Huffman et al. (2000) identified sulfur species emitted
8 by the combustion of several residual fuels oil (RFO) in a fire tube package boiler that is meant
9 to simulate conditions in small institutional and industrial boilers. They found that sulfur was
10 emitted not only as sulfate (26 to 84%), but as thiophenes (13 to 39%) with smaller amounts of
11 sulfides and elemental S. They also found that Ni, V, Fe, Cu, Zn, and Pb are present mainly as
12 sulfates in emissions. Linak et al. (2000) found, when burning RFO, that the fire tube package
13 boiler produced particles with a bimodal size distribution in which about 0.2% of the mass was
14 associated with particles smaller than 0.1 //m AD, with the rest of the mass lying between
15 0.5 and 100 //m. Miller et al. (1998) found that larger particles consisted mainly of cenospheric
16 carbon; whereas trace metals and sulfates were found concentrated in the smaller particles in a
17 fire tube package boiler. In contrast, when RFO was burning in a refractory-lined combustor that
18 is meant to simulate combustion conditions in a large utility residual oil fired boiler, Linak et al.
19 (2000) found that particles were distributed essentially unimodally, with a mean diameter of
20 about 0.1//m.
21 Apart from emissions in the combustion of fossil fuels, trace elements are emitted as the
22 result of various industrial processes such as steel and iron manufacturing and nonferrous metal
23 production (e.g., for Pb, Cu, Ni, Zn, and Cd). As may be expected, emissions factors for the
24 various trace elements are highly source-specific (Nriagu and Pacyna, 1988). Inspection of
25 Table 3D-3 reveals that the emissions from the catalytic cracker and the oil-fired power plant are
26 greatly enriched in rare-earth elements such as La compared to other sources.
27 Emissions from municipal waste incinerators are heavily enriched in Cl, arising mainly
28 from the combustion of plastics and metals that form volatile chlorides. The metals can originate
29 from cans or other metallic objects, and some metals such as Zn and Cd are also additives in
30 plastics or rubber. Many elements such as S, Cl, Zn, Br, Ag, Cd, Sn, In, and Sb are enormously
31 enriched compared to their crustal abundances. A comparison of the trace elemental composition
April 2002 3D-10 DRAFT-DO NOT QUOTE OR CITE
-------
1 of incinerator emissions in Philadelphia, PA (shown in Table 3D-3), with the composition of
2 incinerator emissions in Washington DC, and Chicago, IL (Olmez et al., 1988), shows agreement
3 for most constituents to better than a factor of two.
4 Very limited data exist for characterizing the chemical composition of organic compounds
5 present in particulate emissions from industrial-scale stationary fuel combustion. Oros and
6 Simoneit (2000) have presented the abundance and distribution of organic constituents in coal
7 smokes that have been burned under laboratory conditions. This work provides the basis for
8 further investigation addressing the emissions of coal fired boilers.
9 Rogge et al. (1997a) measured the composition of the organic constituents in the particulate
10 matter emissions from a 50 billion kj/h boiler that was operating at 60% capacity and was
11 burning number 2 distillate fuel oil. The fine carbon particulate matter emissions from this boiler
12 over five tests were composed of an average of 14% organic carbon and 86% elemental carbon
13 (Hildemann et al., 1991). Significant variability in the distribution of organic compounds present
14 in the emissions from two separate tests was observed. Most of the identified organic mass
15 consisted of n-alkanonic acids, aromatic acids, n-alkanes, PAH, oxygeanted PAH, and
16 chlorinated compounds. It is unclear if these emissions are representative of typical fuel oil
17 combustion units in the United States. Rogge et al. (1997b) measured the composition of hot
18 asphalt roofing tar pots, and Rogge et al. (1993b) measured the composition of emissions from
19 home appliances that use natural gas.
20
21 Motor Vehicles
22 Exhaust emissions of particulate matter from gasoline powered motor vehicles and diesel
23 powered vehicles have changed significantly over the past 25 years (Sawyer and Johnson, 1995;
24 Cadle et al., 1999). These changes have resulted from reformulation of fuels, the wide
25 application of exhaust-gas treatment in gasoline-powered motor vehicles, and changes in engine
26 design and operation. Because of these evolving tailpipe emissions, along with the wide
27 variability of emissions between vehicles of the same class (Hildemann et al., 1991; Cadle et al.,
28 1997; Sagebiel et al., 1997; Yanowitz et al., 2000), well-defined average emissions profiles for
29 the major classes of motor vehicles have not been established. Two sampling strategies have
30 been employed to obtain motor vehicle emissions profiles: (1) the measurement of exhaust
31 emissions from vehicles operating on dynamometers and (2) the measurement of integrated
April 2002 3D-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 emissions of motor vehicles driving through roadway tunnels. Dynamometer testing can be used
2 to measure vehicle emissions operating over an integrated driving cycle and allows the
3 measurement of emissions from individual vehicles. However, dynamometer testing requires
4 considerable resources and usually precludes testing a very large number of vehicles. In contrast,
5 a large number of vehicles can be readily sampled in tunnels; however, vehicles driving through
6 tunnels operate over limited driving conditions, and the measurements represent contributions
7 from a large number of vehicle types. As a result, except in a few cases, tunnel tests have not
8 been effective at developing chemically speciated particulate matter emissions profiles for
9 individual motor vehicle classes. Rather, several studies have measured the contribution of both
10 organic and elemental carbon to the particulate matter emissions from different classes of motor
11 vehicles operating on chassis dynamometers.
12 The principal components emitted by diesel and gasoline fueled vehicles are organic carbon
13 (OC) and elemental carbon (EC) as shown in Tables 3D-4a and 4b. As can be seen, the
14 variability among entries for an individual fuel type is large and overlaps that found between
15 different fuel types. On average, the abundance of elemental carbon is larger than that of organic
16 carbon in the exhaust of diesel vehicles; whereas organic carbon is the dominant species in the
17 exhaust of gasoline fueled vehicles. Per vehicle mile, total carbon emissions from light and
18 heavy duty diesel vehicles can range from 1 to 2 orders of magnitude higher than those from
19 gasoline vehicles.
20 As might be expected, most of the PM emitted by motor vehicles is in the PM2 5 size range.
21 Particles in diesel exhaust are typically trimodal (consisting of a nuclei mode, an accumulation
22 mode, and a coarse mode) and are log-normal in form (Kittelson, 1998). More than 90% of the
23 total number of particles are in the nuclei mode, which contains only about 1 to 20% of the
24 particle mass with a mass median diameter of about 0.02 //m; whereas the accumulation mode
25 (with a mass median diameter of about 0.25 //m) contains most of the mass with a smaller
26 fraction (5 to 20%) contained in the coarse mode. Kerminin et al. (1997), Bagley et al. (1998),
27 and Kleeman et al. (2000) also have shown that gasoline and diesel fueled vehicles produce
28 particles that are mostly less than 2.0 //m in diameter. Cadle et al. (1999) found that 91% of PM
29 emitted by in-use gasoline vehicles in the Denver area was in the PM2 5 size range, which
30 increased to 97% for "smokers" (i.e., light-duty vehicles with visible smoke emitted from their
31 tailpipes) and 98% for light-duty diesels. Durbin et al. (1999) found that about 92% of the PM
April 2002 3D-12 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3D-4a. ORGANIC AND ELEMENTAL CARBON FRACTIONS OF DIESEL
AND GASOLINE ENGINE PARTICIPATE MATTER EXHAUST
Heavy-duty diesel engines3
Heavy-duty diesel engines (SPECIATE)b
Light-duty diesel engines0
Light-duty diesel engines (SPECIATE)b
Gasoline engines (hot stabilized)8
Gasoline engines ("smoker" and "high emitter")"'0
Gasoline engines (cold start)"
Organic Carbon
19 ± 8%
21 -36%
30 ± 9%
22 - 43%
56 ± 11%
76 ± 10%
46 ± 14%
Elemental Carbon
75 ± 10%
52 - 54%
61 ± 16%
51-64%
25 ± 15%
7 ± 6%
42 ± 14%
Tujita et al. (1998) and Watson et al. (1998).
bU.S. EPA SPECIATE database.
°Norbeck et al. (1998).
Source: U.S. Environmental Protection Agency (2002).
1 was smaller than 2.5 //m for smokers and diesels. The mass median diameter of the PM emitted
2 by the gasoline vehicles sampled by Cadle et al. (1999) was about 0.12 //m and increased to
3 0.18 //m for smokers and diesels. Corresponding average emissions rates of PM25 found by
4 Cadle et al. (1999) were 552 mg/mile for diesels; 222 mg/mile for gasoline smokers; and
5 38 mg/mile for other gasoline vehicles. The values for gasoline smokers and for diesels appear
6 to be somewhat lower than those given in Table 3D-5; whereas the value for other gasoline
7 vehicles falls in the range given for low and medium gasoline vehicle emissions.
8 Examples of data for the trace elemental composition of the emissions from a number of
9 vehicle classes obtained December 1997 in Colorado, as part of the North Frontal Range Air
10 Quality Study (NFRAQS), are shown in Table 3D-5. As can be seen from Table 3D-5, emissions
11 of total carbon (TC), which is equal to the sum of organic carbon (OC) and elemental carbon
12 (EC), from gasoline vehicles are highly variable. Gillies and Gertler (2000) point out that there is
13 greater variability in the concentrations of trace elements and ionic species than for OC and EC
14 among different source profiles (e.g., SPECIATE, Lawson and Smith [1998], Norbeck et al.
15 [1998]). They suggest that this may arise because emissions of trace elements are not related
16 only to the combustion process, but also to their abundances in different fuels and lubricants and
April 2002 3D-13 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3D-4b. CONTRIBUTION OF ORGANIC CARBON TO PARTICULATE
MATTER CARBON EMISSIONS IN MOTOR VEHICLE EXHAUST COLLECTED
FROM VEHICLES OPERATED ON CHASSIS DYNAMOMETERS
GASOLINE POWERED VEHICLES
Light-duty vehicles
High-CO/VOC-emitting smokers
High-CO/VOC-emitting nonsmokers
Catalyst-equipped vehicles
Noncatalyst vehicles
DIESEL VEHICLES
Light-duty diesel vehicles
Medium-duty diesel vehicles
Heavy-duty diesel vehicles
Heavy-duty diesel vehicles
Year of Tests
1996-97
1994
1994
Mid-1980s
Mid-1980s
1996-1997
1996
1992
Mid-1980s
Test Cycle
FTP
IM-240
IM-240
FTP
FTP
FTP
FTP
c
c
Number of
Vehicles
195a
7
15
7
6
195a
2
6
2
OC % of
Total Carbon
70
91
76
69
89
40
50b
42
45
Notes
A
B
B
C
C
A
D
E
C
Notes:
A. From Cadle et al. (1999). Average of summer and winter cold start emissions.
B. From Sagebiel et al. (1997). Hot start testing of vehicles identified as either high emitters of carbon
monoxide or volatile organic compounds (VOCs).
C. From Hildemann et al. (1991). Cold start tests.
D. From Schauer et al. (1999). Hot start tests of medium duty vehicles operating on an FTP cycle.
E. From Lowenthal et al. (1994). Only includes measurement of vehicles powered by diesel fuel operated
without an exhaust paniculate trap.
aA total of 195 light duty vehicles were tested that include both gasoline powered vehicles and diesel powered
vehicles.
bFraction of paniculate matter consisting of organic carbon was measured with and without an organics denuder
upstream of paniculate filter. Results reported here represent measurement without an organics denuder for
consistency with other measurements. Using an organics denuder, the organic carbon comprised 39% of the
paniculate matter carbon.
Thriving cycle comprised of multiple idle, steady acceleration, constant speed, deceleration steps (see reference
for more details).
1 to wear and tear during vehicle operation. Emissions from gasoline smokers are comparable to
2 those from light-duty diesel vehicles. Thus, older, poorly maintained gasoline vehicles could be
3 significant sources of PM25 (Sagebiel et al., 1997; Lawson and Smith, 1998), in addition to being
April 2002 3D-14 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3D-5. EMISSION RATES (mg/mi) FOR CONSTITUENTS OF PARTICIPATE
MATTER FROM GASOLINE AND DIESEL VEHICLES
Gasoline Vehicles
TC
OC
EC
NO3
SO4=
Na
Mg
Al
Si
P
S
Cl
K
Ca
Fe
Ni
Cu
Zn
Br
Ba
Pb
Low
9.07 ±0.75
6.35 ±0.54
2.72 ±0.52
0.039 ± 0.027
0.158 ±0.036
0.060 ± 0.063
0.036 ±0.022
0.083 ±0.016
0.066 ± 0.008
0.035 ±0.004
0.085 ±0.006
0.024 ±0.012
0.010 ±0.009
0.060 ±0.010
0.143 ±0.004
0.001 ±0.004
0.002 ± 0.004
0.048 ±0.003
0.001 ±0.002
0.013 ±0.136
0.007 ± 0.006
Medium
41. 30 ±1.68
26.02 ±1.31
15.28 ±0.99
0.057 ±0.028
0.518 ±0.043
0.023 ±0.111
0.068 ± 0.027
0.078 ±0.016
0.279 ±0.011
0.152 ±0.007
0.442 ± 0.009
0.038 ±0.012
0.019 ±0.009
0.212 ±0.011
0.756 ±0.005
0.005 ± 0.004
0.016 ±0.003
0.251 ±0.004
0.016 ±0.002
0.009 ±0.138
0.085 ± 0.005
High
207.44 ± 7.29
95.25 ±4.28
112.19 ±5. 82
0.141 ±0.031
0.651 ±0.052
0.052 ±0.092
0.041 ±0.033
0.057 ±0.014
0.714 ±0.012
0.1 13 ±0.007
0.822 ± 0.022
0.081 ±0.020
0.031 ±0.035
0.210 ±0.030
1.047 ±0.010
0.011 ±0.005
0.021 ±0.005
0.265 ± 0.023
0.079 ±0.003
0.011 ±0.299
0.255 ±0.008
Smoker
456.38 ±16.80
350.24 ±15.27
106. 14 ±5.42
0.964 ±0.051
2.160 ±0.137
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
2.515 ±0.116
0.140±0.117
0.033 ±0.386
0.362 ±0.250
2.438 ±0.054
0.008 ±0.017
0.071 ±0.018
0.188 ±0.272
0.047 ±0.012
0.380 ±2.175
0.345 ±0.032
Diesel
Light Duty
373.43 ±13.75
132.01 ±5.82
241.42 ±12.11
1.474 ±0.071
2.902 ±0.165
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
0.000 ± 0.000
2.458 ±0.124
0.228 ±0.114
0.000 ± 0.426
0.150 ±0.304
0.515 ±0.057
0.014 ±0.018
0.024 ±0.021
0.000 ±0.299
0.003 ±0.014
0.428 ±2.390
0.153 ±0.033
Vehicles
Heavy Duty
1570.69 ±58.24
253. 94 ±16.12
1316.75 ±55. 33
1.833 ±1.285
3.830 ±1.286
1.288 ±2.160
1.061 ±0.729
0.321 ±0.543
8.018 ±0.221
0.407 ±0.136
3.717±0.111
0.881 ±0.221
0.064 ± 0.248
0.716 ±0.107
0.376 ±0.055
0.002 ±0.057
0.001 ±0.062
0.707 ±0.032
0.012 ±0.050
0.493 ±3. 108
0.008 ±0.154
Source: Lawson and Smith (1998).
1 significant sources of gaseous pollutants (e.g., Calvert et al., 1993). Durbin et al. (1999) point
2 out that although smokers constitute only 1.1 to 1.7% of the light-duty fleet in the South Coast
3 Air Quality Management District in California, they contribute roughly 20% of the total PM
April 2002
3D-15
DRAFT-DO NOT QUOTE OR CITE
-------
1 emissions from the light-duty fleet. In general, motor vehicles that are high emitters of
2 hydrocarbons and carbon monoxide also will tend to be high emitters of PM (Sagebiel
3 et al.,1997; Cadle et al., 1997). Particle emission rates, even in newer vehicles, also are
4 correlated with vehicle acceleration; and emissions occur predominantly during periods of heavy
5 acceleration (Maricq et al., 1999).
6 Although the data shown in Table 3D-5 indicate that S (mainly in the form of sulfate) is a
7 minor component of PM2 5 emissions, S may be the major component of the ultrafine particles
8 that are emitted by either diesel or internal combustion engines (Gertler et al., 2000). It is not
9 clear what the source of the small amount of Pb seen in the auto exhaust profile is. It is
10 extremely difficult to find suitable tracers for automotive exhaust because Pb has been removed
11 from gasoline. However, it also should be remembered that restrictions in the use of leaded
12 gasoline have resulted in a dramatic lowering of ambient Pb levels.
13 Several tunnel studies have measured the distribution of organic and elemental carbon in
14 the integrated exhaust of motor vehicle fleets comprising several classes of motor vehicles
15 (Pierson and Brachaczek, 1983; Weingartner et al., 1997a; Fraser et al., 1998a). The study by
16 Fraser et al. (1998a) found that organic carbon constituted 46% of the carbonaceous PM
17 emissions from the vehicles operating in the Van Nuys tunnel in Southern California in the
18 Summer of 1993. Although diesel vehicles constituted only 2.8% of the vehicles measured by
19 Fraser et al. (1998a), the contribution of the organic carbon to the total paniculate carbon
20 emissions obtained in the Van Nuys tunnels is in reasonable agreement with the dynamometer
21 measurements shown in Table 3D-4b.
22 Very few studies have reported comprehensive analyses of the organic composition of
23 motor vehicle exhaust. The measurements by Rogge et al. (1993c) are the most comprehensive
24 but are not expected to be the best representation of current motor vehicle emissions because
25 these measurements were made in the mid-1980s. Measurements reported by Fraser et al. (1999)
26 were made in a tunnel study conducted in 1993 and represent integrated diesel and gasoline
27 powered vehicle emissions. In addition, exhaust emissions from two medium-duty diesel
28 vehicles operating over an FTP cycle were analyzed by Schauer et al. (1999). A unique feature
29 of both the measurements by Faser et al. (1999) and Schauer et al. (1999) is that they include the
30 quantification of unresolved complex mixture (UCM), which comprises aliphatic and cyclic
31 hydrocarbons that cannot be resolved by gas chromatography (Schauer et al., 1999). Schauer
April 2002 3D-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 et al. (1999) have shown that all of the organic compound mass in their diesel exhaust samples
2 could be extracted and eluted by CG/MS techniques even though not all of the organic compound
3 mass can identified on a single compound basis
4 motor vehicle exhaust
5
TABLE 3D-6.
Source
Gasoline and diesel-
powered vehicles
driving through the
Van Nuys Tunnel
(Eraser etal, 1999)a
Medium-duty diesel
vehicles operated over
an FTP Cycle
(Schaueretal., 1999)
measured by Fraser et al.
. Table 3D-6 summarizes the composition of
(1999) and Schauer
etal. (1999).
SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
EMITTED FROM
Compound Class
n-Alkanes
Petroleum biomarkers
PAH
Aromatic acids
Aliphatic acids
Substituted aromatic
UCMb
n-Alkanes
Petroleum biomarkers
PAH
Aliphatic acids
Aromatic acids
Saturated cycloalkanes
UCMb
MOTOR VEHICLES
Contribution to
Paniculate Mass (%)
0.009
0.078
0.38
0.29
0.21
0.042
23.0
0.22
0.027
0.54
0.24
0.014
0.037
22.2
Dominant Contributors to
Emissions of Compound Class
C21 through C29
Hopanes and steranes
No dominant compound
Benzenedicarboxylic acids
Palmitic and stearic acids
No dominant compound
C20 through C28
Hopanes and steranes
No dominant compound
n-Octadecanoic acid
Methylbenzoic acid
C21 through C25
"Includes emissions of brake wear, tire wear, and resuspension of road dust associated with motor vehicle traffic.
bUnresolved complex mixture.
1 Several studies have measured the distribution of polycyclic aromatic hydrocarbons (PAHs)
2 in motor vehicles exhaust from on-road vehicles (Westerholm et al., 1991; Lowenthal et al.,
3 1994; Venkataraman et al., 1994; Westerholm and Egeback, 1994; Reilly et al., 1998; Cadle
April 2002
3D-17
DRAFT-DO NOT QUOTE OR CITE
-------
1 et al., 1999, Weingartner et al., 1997b; Marr et al., 1999). Cadle et al. (1999) found high
2 molecular weight PAHs (PAHs with molecular weights greater than or equal to 202 g/mole)
3 constitute 0.1 to 7.0% of the particulate matter emissions from gasoline powered and diesel
4 powered light duty vehicles. It is important to note, however, that PAHs with molecular weights
5 of 202 (fluoranthene, acephenanthrylene, and pyrene), 226 (benzo[ghi]fluoranthene and
6 cyclopenta[cd]pyrene), and 228 (benz[a]anthracene, chrysene, and triphenylene) exist in both the
7 gas-phase and particle-phase at atmospheric conditions (Fraser et al., 1998b) although those with
8 molecular weight of 228 are predominantly associated with particles, with only traces in the
9 gas-phase (Arey et al., 1987). Excluding these semivolatile PAHs, the contribution of
10 nonvolatile PAHs to the parti culate matter emitted from the light-duty vehicles sampled by Cadle
11 et al. (1999) ranges from 0.013 to 0.18%. These measurements are in good agreement with the
12 tunnel study conducted by Fraser et al. (1999) and the heavy-duty diesel truck and bus exhaust
13 measurements by Lowenthal et al. (1994), except that the nonvolatile PAH emissions from the
14 heavy duty diesel vehicles tested by Lowenthal et al. (1994) were moderately higher, making up
15 approximately 0.30% of the particulate matter mass emissions.
16
17 Biomass Burning
18 In contrast to the mobile and stationary sources discussed earlier, emissions from biomass
19 burning in wood stoves and forest fires are strongly seasonal and can be highly episodic within
20 their peak emissions seasons. The burning of fuelwood is confined mainly to the winter months
21 and is acknowledged to be a major source of ambient air particulate matter in the northwestern
22 United States during the heating season. Forest fires occur primarily during the driest seasons of
23 the year in different areas of the country and are especially prevalent during prolonged droughts.
24 PM produced by biomass burning outside the United States (e.g., in Central America during the
25 spring of 1988) also can affect ambient air quality in the United States.
26 An example of the composition of fine particles (PM25) produced by wood stoves is shown
27 in Figure 3D-4. These data were obtained in Denver during the winter of 1987-1988 (Watson
28 and Chow, 1994). As was the case for motor vehicle emissions, organic and elemental carbon
29 are the major components of particulate emissions from wood burning. It should be remembered
30 that the relative amounts shown for organic carbon and elemental carbon vary with the type of
31 stove, the stage of combustion, and the type and condition of the fuelwood. Fine particles are
April 2002 3D-18 DRAFT-DO NOT QUOTE OR CITE
-------
10"
10
^4^
Chemical Species
Figure 3D-4. Chemical abundances for PM2 5 emissions from wood burning in Denver,
CO. Solid bars represent fractional abundances, and the error bars
represent variability in species abundances. Error bars represent detection
limits when there are no solid bars.
Source: Watson and Chow (1994).
1
2
3
4
5
9
10
11
12
dominant in smoke studies of wood burning emissions. For instance, the mass median diameter
of wood particles was found to be about 0.17 //m in a study of the emissions from burning
hardwood, softwood, and synthetic logs (Dasch, 1982).
Kleeman et al. (1999) showed that the particles emitted by the combustion of wood in
fireplaces are predominately less than 1.0 //m in diameter, such that the composition of fine PM
(PM2 5) emitted from fireplace combustion of wood is representative of the total particulate
matter emissions from this source. Hildemann et al. (1991) and McDonald et al. (2000) reported
that smoke from fireplace and wood stove combustion consists of 48% to 71% OC and 2.9% to
15% EC. Average elemental and organic carbon contents for these measurements are shown in
Table 3D-7. It should be noted that the two methods used for the measurements shown in
Table 3D-7 have been reported to produce different relative amounts of OC and EC for wood
April 2002
3D-19
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3D-7. MASS EMISSIONS, ORGANIC CARBON, AND ELEMENTAL
CARBON EMISSIONS FROM RESIDENTIAL COMBUSTION OF WOOD
Wood Type
Softwood
Softwood
Hardwood
Hardwood
Hardwood
Combustion
Type
Fireplace
Fireplace
Fireplace
Fireplace
Wood Stove
Average Mass
Emission Rate
(gkg"1 of wood
burned)
13.0
5.14
5.28
5.66
3.96
Number
of Tests
2
5
3
5
8
Percent
Organic
Carbona
48.4
58.5
48.4
63.2
71.2
Percent
Elemental
Carbon8
5.2
15.0
2.9
7.0
9.0
References
Hildemann et al. (1991)
McDonald et al. (2000)
Hildemann etal. (1991)
McDonald et al. (2000)
McDonald et al. (2000)
"Hildemann et al. (1991) used the method described by Birch and Gary (1996) to measure EC and McDonald
et al. (2000) used the method reported by Chow et al. (1993) to measure OC.
1 smoke samples but show good agreement for total carbon (OC + EC) measurements (Chow
2 etal., 1993).
3 Hawthorne et al. (1988) and Hawthorne et al. (1989) measured gas-phase and particle-
4 phase derivatives of guaiacol (2-methoxyphenol), syringol (2,6-dimethoxyphenol), phenol, and
5 catechol (1,2-benzenediol) in the downwind plume of 28 residential wood stoves and fireplaces.
6 Rogge et al. (1998) reported a broad range of particle-phase organic compounds in the wood
7 smoke samples collected by Hildemann et al. (1991), which include n-alkanes, n-alkanoic acids,
8 n-alkenoic acids, dicarboxylic acids, resin acids, phytosterols, polycyclic aromatic hydrocarbons
9 (PAH), and the compounds reported by Hawthorne et al. (1989). Supplementing these
10 measurements, McDonald et al. (2000) reported the combined gas-phase and particle-phase
11 emissions of PAH and the compounds quantified by Hawthorne et al. (1989). The measurements
12 by Rogge et al. (1998), which represent a comprehensive data set of the organic compounds
13 present in wood smoke aerosol, are summarized in Table 3D-8. It should be noted, however, that
14 these nearly 200 compounds account for only approximately 15 to 25% of the organic carbon
15 particle mass emitted from the residential combustion of wood. Simoneit et al. (1999) have
16 shown that levoglucosan constitutes a noticeable portion of the organic compound mass not
17 identified by Rogge et al. (1998). In addition, Elias et al. (1999) used high-temperature gas
18 chromatography/mass spectrometry (HTGC-MS) to measure high-molecular-weight organic
April 2002
3D-20
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3D-8. SUMMARY OF PARTICLE-PHASE ORGANIC COMPOUNDS
EMITTED FROM THE COMBUSTION OF WOOD IN FIREPLACES*
Biomass Type
Fireplace
combustion of
softwood
Fireplace
combustion of
hardwood
Contribution to Paniculate Dominant Contributors to Emissions
Compound Class Mass (%) of Compound Class
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Dicarboxylic acids
Resin acids
Substituted phenols
Phytosterols
PAH
Oxygenated PAH
n-Alkanes
n-Alkanoic acids
n-Alkenoic acids
Dicarboxylic acids
Resin acids
Substituted phenols
Phytosterols
PAH
Oxygenated PAH
0.039
0.45
0.12
0.36
1.28
3.30
0.37
0.092
0.019
0.044
1.33
0.049
0.42
0.11
8.23
0.21
0.13
0.020
C21 through C31
c c c c c c
*"16! M8> ^205 *^21> ^225 ^24
Oleic and linoleic acid
Malonic acid
Abietic, dehydroabietic, isopimaric,
pimaric, and sandaracopimaric acids
Benzenediols and guaiacols
-Sitosterol
Fluoranthene and pyrene
IH-phenalen-l-one
C21 through C29
Cis, C22, C24, C26
Oleic and linoleic acid
Succinic acid
Dehydroabietic acid
Benzediols, guaiacols, and syringols
-sitosterol
No dominant compounds
IH-phenalen-l-one
*Note: Measurements were made using a dilution sampler and no semivolatile organic compound sorbent.
Source: Rogge et al. (1998).
1 compounds in smoke from South American leaf and stem litter biomass burning. These
2 compounds cannot be measured by the analytical techniques employed by Rogge et al. (1998)
3 and, therefore, are strong candidates to make up some of the unidentified organic mass in the
4 wood smoke samples analyzed by Rogge et al. (1998). These compounds, which include
April 2002
3D-21
DRAFT-DO NOT QUOTE OR CITE
-------
1 triterpenyl fatty acid esters, wax esters, triglycerides, and high-molecular-weight n-alkan-2-ones,
2 are expected to be present in North American biomass smoke originating from agricultural
3 burning, forest fires, grassland fires, and wood stove/fireplace smoke.
4 Measurements of aerosol composition, size distributions, and aerosol emissions factors
5 have been made in biomass burning plumes, either on towers (Susott et al., 1991) or aloft on
6 fixed-wing aircraft (e.g., Radke et al., 1991) or on helicopters (e.g., Cofer et al., 1988). As was
7 found for wood stove emissions, the composition of biomass burning emissions is strongly
8 dependent on the stage of combustion (i.e., flaming, smoldering, or mixed), and the type of
9 vegetation (e.g., forest, grassland, scrub). Over 90% of the dry mass in particulate biomass
10 burning emissions is composed of organic carbon (Mazurek et al., 1991). Ratios of organic
11 carbon to elemental carbon are highly variable, ranging from 10:1 to 95:1, with the highest ratio
12 found for smoldering conditions and the lowest for flaming conditions. Emissions factors for
13 total particulate emissions increase by factors of two to four in going from flaming to smoldering
14 stages in the individual fires studied by Susott et al. (1991).
15 Particles in biomass burning plumes from a number of different fires were found to have
16 three distinguishable size modes: (1) a nucleation mode, (2) an accumulation mode, and
17 (3) a coarse mode (Radke et al., 1991). Based on an average of 81 samples, approximately 70%
18 of the mass was found in particles <3.5 //m in aerodynamic diameter. The fine particle
19 composition was found to be dominated by tarlike, condensed hydrocarbons; and the particles
20 were usually spherical in shape. Additional information for the size distribution of particles
21 produced by vegetation burning is shown in Figure 3D-2.
22 An example of ambient data for the composition of PM2 5 collected at a tropical site that
23 was heavily affected by biomass burning is shown in Table 3D-9. The samples were collected
24 during November of 1997 on the campus of Sriwijaya University, which is located in a rural
25 setting on the island of Sumatra in Indonesia (Pinto et al., 1998). The site was subjected
26 routinely to levels of PM25 well in excess of the U.S. NAAQS as a result of the Indonesian
27 biomass fires from the summer of 1997 through the spring of 1998. As can be seen from a
28 comparison of the data shown in Table 3D-9 with those shown in Figure 3D-4, there are a
29 number of similarities and differences (especially with regard to the heavy metal content) in the
30 abundances of many species. The abundances of some crustal elements (e.g., Si, Fe) are higher
31
April 2002 3D-22 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 3D-9. MEAN AEROSOL COMPOSITION AT TROPICAL SITE
(SRIWIJAYA UNIVERSITY, SUMATRA, INDONESIA) AFFECTED
HEAVILY BY BIOMASS BURNING EMISSIONS3
Component
OC
EC
s
-------
1 Natural Sources
2 Although sea-salt aerosol production is confined to salt water bodies, it is included here
3 because many marine aerosols can exert a strong influence on the composition of the ambient
4 aerosol in coastal areas. In some respects, the production of sea-salt aerosols is like that of
5 windblown dust, in that both are produced by wind agitation of the surface. The difference
6 between the two categories arises because sea-salt particles are produced from the bursting of air
7 bubbles rising to the sea surface. Air bubbles are formed by the entrainment of air into the water
8 by breaking waves. The surface energy of a collapsing bubble is converted to kinetic energy in
9 the form of a j et of water that can ej ect drops above the sea surface. The mean diameter of the j et
10 drops is about 15% of the bubble diameter (Wu, 1979). Bubbles in breaking waves range in size
11 from a few //m to several mm in diameter. Field measurements by Johnson and Cooke (1979) of
12 bubble size spectra show maxima in diameters at around 100 //m, with the bubble size
13 distribution varying as (d/d0)"5 with d0 = 100 //m.
14 Because sea-salt particles receive water from the surface layer, which is enriched in organic
15 compounds, aerosol drops are composed of this organic material in addition to sea salt (about
16 3.5% by weight in seawater). Na+ (30.7%), Cl' (55.0%), S
-------
1 animal debris, such as insect fragments, also are to be found in ambient aerosol samples in this
2 size range. Although material from all the foregoing categories may exist as individual particles,
3 bacteria usually are found attached to other dust particles (Warneck, 1988). Smaller bioaerosol
4 particles include viruses, individual bacteria, protozoa, and algae (Matthias-Maser and Jaenicke,
5 1994). In addition to natural sources, other sources of bioaerosol include industry (e.g., textile
6 mills), agriculture, and municipal waste disposal (Spendlove, 1974). The size distribution of
7 bioaerosols has not been characterized as well as it has for other categories of airborne particles.
8 Trace metals are emitted to the atmosphere from a variety of sources such as sea spray,
9 wind-blown dust, volcanoes, wildfires and biotic sources (Nriagu, 1989). Biologically mediated
10 volatilization processes (e.g., biomethylation) are estimated to account for 30 to 50% of the
11 worldwide total Hg, As, and Se emitted annually; whereas other metals are derived principally
12 from pollens, spores, waxes, plant fragments, fungi, and algae. It is not clear, however, how
13 much of the biomethylated species are remobilized from anthropogenic inputs. Median ratios of
14 the natural contribution to globally averaged total sources for trace metals are estimated to be
15 0.39 (As), 0.15 (Cd), 0.59 (Cr), 0.44 (Cu), 0.41 (Hg), 0.35 (Ni), 0.04 (Pb), 0.41 (Sb), 0.58 (Se),
16 0.25 (V), and 0.34 (Zn), suggesting a significant natural source for many trace elements.
17 It should be noted, however, that these estimates are based on emissions estimates that have
18 uncertainty ranges of an order of magnitude.
April 2002 3D-25 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Arey, J.; Zielinska, B.; Atkinson, R.; Winer, A. M. (1987) Polycyclic aromatic hydrocarbon and nitroarene
3 concentrations in ambient air during a wintertime high-NO^ episode in the Los Angeles basin.
4 Atmos. Environ. 21: 1437-1444.
5 Bagley, S. T.; Gratz, L. D.; Johnson, J. H.; McDonald J. F. (1998) Effects of an oxidation catalytic converter and a
6 biodiesel fuel on the chemical, mutagenic, and particle size characteristics of emissions from a diesel engine.
7 Environ. Sci. Technol. 32: 1183-1191.
8 Birch, M. E.; Gary, R. A. (1996) Elemental carbon-based method for occupational monitoring of paniculate diesel
9 exhaust: methodology and exposure issues. Analyst 121: 1183-1190.
10 Cadle, S. H.; Mulawa, P. H.; Ball, J.; Donase, C; Weibel, A.; Sagebiel, J. C; Knapp, K. T.; Snow, R. (1997)
11 Paniculate emission rates from in-use high-emitting vehicles recruited in Orange county, California. Environ.
12 Sci. Technol. 31:3405-3412.
13 Cadle, S. H.; Mulawa, P.; Hunsanger, E. C.; Nelson, K.; Ragazzi, R. A.; Barrett, R.; Gallagher, G. L.; Lawson,
14 D. R.; Knapp, K. T.; Snow, R. (1999) Light-duty motor vehicle exhaust paniculate matter measurement in the
15 Denver, Colorado, area. J. Air Waste Manage. Assoc. 49: PM-164-174.
16 Calvert, J. G.; Heywood, J. B.; Sawyer, R. F.; Seinfeld, J. H. (1993) Achieving acceptable air quality: some
17 reflections on controlling vehicle emissions. Science (Washington, DC) 261: 37-45.
18 Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993) The DRI
19 thermal/optical reflectance carbon analysis system: description, evaluation and applications in U.S. air quality
20 studies. In: Lodge, J. P., Jr. Fourth international conference on carbonaceous particles in the atmosphere;
21 April 1991; Vienna, Austria. Atmos. Environ. Part A 27: 1185-1201.
22 Chow, J. C.; Watson, J. G.; Houck, J. E.; Pritchett, L. C.; Rogers, C. F.; Frazier, C. A.; Egami, R. T.; Ball, B. M.
23 (1994) A laboratory resuspension chamber to measure fugitive dust size distributions and chemical
24 compositions. Atmos. Environ. 28: 3463-3481.
25 Cofer, W. R., Ill; Levine, J. S.; Sebacher, D. I.; Winstead, E. L.; Riggin, P. J.; Brass, J. A.; Ambrosia, V. G. (1988)
26 Paniculate emissions from a mid-latitude prescribed chaparral fire. J. Geophys. Res. [Atmos.] 93: 5207-5212.
27 Dasch, J. M. (1982) Paniculate and gaseous emissions from wood-burning fireplaces. Environ. Sci. Technol.
28 16:639-645.
29 Durbin, T. D.; Smith, M. R.; Norbeck, J. M.; Truex, T. J. (1999) Population density, paniculate emission
3 0 characterization, and impact on the paniculate inventory of smoking vehicles in the South Coast Air Quality
31 Management District. J. Air Waste Manage. Asssoc. 49: 28-38.
32 Elias, V. O.; Simoneit, B. R. T.; Pereira, A. S.; Cabral, J. A.; Cardoso, J. N. (1999) Detection of high molecular
3 3 weight organic tracers in vegetation smoke samples by high-temperature gas chromatography-mass
34 spectrometry. Environ. Sci. Technol. 33: 2369-2376.
35 Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. (1998a) Gas-phase and particle-phase organic compounds emitted
36 from motor vehicle traffic in a Los Angeles roadway tunnel. Environ. Sci. Technol. 32: 2051-2060.
37 Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T.; Rasmussen, R. A. (1998b) Air quality model evaluation data for
38 organics. 5. C6-C22 nonpolar and semipolar aromatic compounds. Environ. Sci. Technol. 32: 1760-1770.
39 Fraser, M. P.; Cass, G. R.; Simoneit, B. R. T. (1999) Paniculate organic compounds emitted from motor vehicle
40 exhaust and in the urban atmosphere. Atmos. Environ. 33: 2715-2724.
41 Fujita, E.; Watson, J. G.; Chow, J. C.; Robinson, N. F.; Richards, L. W.; Kumar, N. (1998) Northern front range air
42 quality study, volume C: source apportionment and simulation methods and evaluation. Reno, NV: Desert
43 Research Institute; prepared for Colorado State University, Cooperative Institute for Research in the
44 Atmosphere.
45 Gertler, A. W.; Abu-Allaban, M.; Coulombe, W.; Gilles, J. A.; Pierson, W. R.; Rogers, C. F.; Sagebiel, J. C.;
46 Tarnay, L. (2000) Preliminary results of a tunnel study to characterize mobile source paniculate emissions.
47 Presented at: PM2000: paniculate matter and health—the scientific basis for regulatory decision-making,
48 specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management
49 Association.
50 Gillies, J. A.; Gertler, A. W. (2000) Comparison and evaluation of chemically speciated mobile source PM2 5
51 paniculate matter profiles. J. Air Waste Manage. Assoc. 50: 1459-1480.
52 Hawthorne, S. B.; Miller, D. J.; Barkley, R. M.; Krieger, M. S. (1988) Identification of methoxylated phenols as
53 candidate tracers for atmospheric wood smoke pollution. Environ. Sci. Technol. 22: 1191-1196.
April 2002 3D-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hawthorne, S. B.; Krieger, M. S.; Miller, D. J.; Mathiason, M. B. (1989) Collection and quantitation of
2 methoxylated phenol tracers for atmospheric pollution from residential wood stoves. Environ. Sci. Technol.
3 23:470-475.
4 Hertel, O.; Christensen, J.; Hov, 0. (1994) Modelling of the end products of the chemical decomposition of DMS in
5 the marine boundary layer. Atmos. Environ. 28: 2431-2449.
6 Hildemann, L. M.; Markowski, G. R.; Jones, M. C.; Cass, G. R. (1991) Submicrometer aerosol mass distributions of
7 emissions from boilers, fireplaces, automobiles, diesel trucks, and meat-cooking operations. Aerosol Sci.
8 Technol. 14: 138-152.
9 Hopke, P. K.; Lamb, R. E.; Natusch, D. F. S. (1980) Multielemental characterization of urban roadway dust.
10 Environ. Sci. Technol. 14: 164-172.
11 Houck, J. E.; Chow, J. C.; Watson, J. G.; Simons, C. A.; Pritchett, L. C.; Goulet, J. M.; Frazier, C. A. (1989)
12 Determination of particle size distribution and chemical composition of paniculate matter from selected
13 sources in California: volume I and executive summary (final report). Sacramento, CA: California Air
14 Resources Board.
15 Houck, J. E.; Goulet, J. M.; Chow, J. C.; Watson, J. G.; Pritchett, L. C. (1990) Chemical characterization of
16 emission sources contributing to light extinction. In: Mathai, C. V., ed. Visibility and fine particles: an
17 A&WMA/EPA international specialty conference; October 1989; Estes Park, CO. Pittsburgh, PA: Air &
18 Waste Management Association; pp. 437-446. (A&WMA transactions series no. TR-17).
19 Huffman, G. P.; Huggins, F. E.; Shah, N.; Huggins, R.; Linak, W. P.; Miller, C. A.; Pugmire, R. J.; Meuzelaar,
20 H. L. C.; Seehra, M. S.; Manivannan, A. (2000) Characterization of fine paniculate matter produced by
21 combustion of residual fuel oil. J. Air Waste Manage. Assoc. 50: 1106-1114.
22 Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. (1996) Particle concentrations, gas-particle partitioning,
23 and species intercorrelations for polycyclic aromatic hydrocarbons (PAH) emitted during biomass burning.
24 Atmos. Environ. 30: 3825-3835.
25 Johnson, B. D.; Cooke, R. C. (1979) Bubble populations and spectra in coastal waters: a photographic approach.
26 J. Geophys. Res. C: Oceans Atmos. 84: 3761-3766.
27 Kerminen, V. M.; Makela, T. E.; Ojanen, C. H.; Hillamo, R. E.; Vilhunen, J. K.; Rantanen, L.; Havers, N.;
28 Von Bohlen, A.; Klockow, D. (1997) Characterization of the paniculate phase in the exhaust from a diesel
29 car. Environ. Sci. Technol. 31: 1883-1889.
30 Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
31 Kleeman, M. J.; Schauer, J. J.; Cass, G. R. (1999) Size and composition distribution of fine paniculate matter
32 emitted from wood burning, meat charbroiling, and cigarettes. Environ. Sci. Technol. 33: 3516-3523.
33 Kleeman, M. J.; Schauer, J. J.; Cass, G. R. (2000) Size and composition distribution of fine paniculate matter
34 emitted from motor vehicles. Environ. Sci. Technol. 34: 1132-1142.
35 Lawson, D. R.; Smith, R. E. (1998) The northern front range air quality study: a report to the Governor and General
36 Assembly. Fort Collins, CO: Colorado State University; December.
37 Linak, W. P.; Miller, C. A.; Wendt, J. O. L. (2000) Comparison of particle size distributions and elemental
3 8 partitioning from the combustion of pulverized coal and residual fuel oil. J. Air Waste Manage. Assoc.
39 50: 1532-1544.
40 Lowenthal, D. H.; Zielinska, B.; Chow, J. C.; Watson, J. G.; Gautam, M.; Ferguson, D. H.; Neuroth, G. R.; Stevens,
41 K. D. (1994) Characterization of heavy-duty diesel vehicle emissions. Atmos. Environ. 28: 731-743.
42 Maricq, M. M.; Podsiadlik, D. H.; Chase, R. E. (1999) Gasoline vehicle particle size distributions: comparison of
43 steady state, FTP, and US06 measurements. Environ. Sci. Technol. 33: 2007-2015.
44 Marr, L. C.; Kirchstetter, T. W.; Harley, R. A.; Miguel, A. H.; Hering, S. V; Hammond, S. K. (1999)
45 Characterization of polycyclic aromatic hydrocarbons in motor vehicle fuels and exhaust emissions.
46 Environ. Sci. Technol. 33: 3091-3099.
47 Mason, B. (1966) Principles of geochemistry. 3rd ed. New York, NY: John Wiley & Sons, Inc.
48 Matthias-Maser, S.; Jaenicke, R. (1994) Examination of atmospheric bioaerosol particles with radii > 0.2 /j.m.
49 J. Aerosol Sci. 25: 1605-1613.
50 Mazurek, M. A.; Cofer, W. R., Ill; Levine, J. S. (1991) Carbonaceous aerosols from prescribed burning of a boreal
51 forest ecosystem. In: Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric
52 implications. Cambridge, MA: MIT Press; pp. 258-263.
53 McDonald, J. D.; Zielinska, B.; Fujita, E. M.; Sagebiel, J. C.; Chow, J. C.; Watson, J. G. (2000) Fine particle and
54 gaseous emission rates from residential wood combustion. Environ. Sci. Technol. 34: 2080-2091.
April 2002 3D-27 DRAFT-DO NOT QUOTE OR CITE
-------
1 Miller, C. A.; Linak, W. P.; King, C.; Wendt, J. O. L. (1998) Fine particle emissions from heavy fuel oil combustion
2 in a firetube package boiler. Combust. Sci. Technol. 134: 477-502.
3 Norbeck, J. M; Durbin, T. D.; Truex, T. J. (1998) Measurement of primary paniculate matter emissions from
4 light-duty motor vehicles. Riverside, CA: University of California, College of Engineering, Center for
5 Environmental Research and Technology; prepared for Coordinating Research Council, Inc. and South Coast
6 Air Quality Management District.
7 Nriagu, J. O. (1989) A global assessment of natural sources of atmospheric trace metals. Nature (London)
8 338:47-49.
9 Nriagu, J. O.; Pacyna, J. M. (1988) Quantitative assessment of worldwide contamination of air, water and soils by
10 trace metals. Nature (London) 333: 134-139.
11 Olmez, I.; Sheffield, A. E.; Gordon, G. E.; Houck, J. E.; Pritchett, L. C.; Cooper, J. A.; Dzubay, T. G.; Bennett,
12 R. L. (1988) Compositions of particles from selected sources in Philadelphia for receptor modeling
13 applications. JAPCA 38: 1392-1402.
14 Oros, D. R.; Simoneit, B. R. T. (2000) Identification and emission rates of molecular tracers in coal smoke
15 paniculate matter. Fuel 79: 515-536.
16 Pierson, W. R.; Brachaczek, W. W. (1983) Paniculate matter associated with vehicles on the road. II. Aerosol Sci.
17 Technol. 2: 1-40.
18 Pinto, J. P.; Grant, L. D.; Hartlage, T. A. (1998) Report on U.S. EPA air monitoring of haze from S.E. Asia biomass
19 fires. Research Triangle Park, NC: U.S. Environmental Protection Agency, National Center for
20 Environmental Assessment; report no. EPA/600/R-98/071.
21 Radke, L. F.; Hegg, D. A.; Hobbs, P. V.; Nance, J. D.; Lyons, J. H.; Laursen, K. K.; Weiss, R. E.; Riggan, P. J.;
22 Ward, D. E. (1991) Paniculate and trace gas emissions from large biomass fires in North America. In:
23 Levine, J. S., ed. Global biomass burning: atmospheric, climatic, and biospheric implications. Cambridge,
24 MA: MIT Press; pp. 209-224.
25 Reilly, P. T. A.; Gieray, R. A.; Whitten, W. B.; Ramsey, J. M. (1998) Real-time characterization of the organic
26 composition and size of individual diesel engine smoke particles. Environ. Sci. Technol. 32: 2672-2679.
27 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1991) Sources of fine organic
28 aerosol. 1. Charbroilers and meat cooking operations. Environ. Sci. Technol. 25: 1112-1125.
29 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993a) Sources of fine organic
30 aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks.
31 Environ. Sci. Technol. 27: 1892-1904.
32 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
33 aerosol. 5. Natural gas home appliances. Environ. Sci. Technol. 27: 2736-2744.
34 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993c) Sources of fine organic
35 aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks. Environ. Sci.
36 Technol. 27:636-651.
37 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993d) Sources of fine organic
3 8 aerosol. 4. Paniculate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol.
39 27:2700-2711.
40 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1994) Sources of fine organic
41 aerosol. 6. Cigarette smoke in the urban atmosphere. Environ. Sci. Technol. 28: 1375-1388.
42 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997a) Sources of fine organic
43 aerosol. 8. Boilers burning no. 2 distillate fuel oil. Environ. Sci. Technol. 31: 2731-2737.
44 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1997b) Sources of fine organic
45 aerosol. 7. Hot asphalt roofing tar pot fumes. Environ. Sci. Technol. 31: 2726-2730.
46 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1998) Sources of fine organic
47 aerosol. 9. Pine, oak, and synthetic log combustion in residential fireplaces. Environ. Sci. Technol. 32: 13-22.
48 Sagebiel, J. C.; Zielinska, B.; Walsh, P. A.; Chow, J. C.; Cadle, S. H.; Mulawa, P. A.; Knapp, K. T.; Zweidinger,
49 R. B.; Snow, R. (1997) PM-10 exhaust samples collected during IM-240 dynamometer tests of in-service
50 vehicles in Nevada. Environ. Sci. Technol. 31: 75-83.
51 Savoie, D. L.; Prospero, J. M. (1982) Particle size distribution of nitrate and sulfate in the marine atmosphere.
52 Geophys. Res. Lett. 9: 1207-1210.
53 Sawyer, R. F.; Johnson, J. H. (1995) Diesel emissions and control technology. Diesel exhaust: a critical analysis of
54 emissions, exposure, and health effects. Cambridge, MA: Health Effects Institute.
April 2002 3D-28 DRAFT-DO NOT QUOTE OR CITE
-------
1 Schauer, J. J.; Cass, G. R. (2000) Source apportionment of wintertime gas-phase and particle-phase air pollutants
2 using organic compounds as tracers. Environ. Sci. Technol. 34: 1821-1832.
3 Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T. (1999) Measurement of emissions from air pollution
4 sources. 2. Q through C30 organic compounds from medium duty diesel trucks. Environ. Sci. Technol.
5 33: 1578-1587.
6 Simoneit, B. R. T. (1985) Application of molecular marker analysis to vehicular exhaust for source reconciliations.
7 Int. J. Environ. Anal. Chem. 22: 203-233.
8 Simoneit, B. R. T.; Schauer, J. J.; Nolle, C. G.; Oros, D. R.; Elias, V. O.; Fraser, M. P.; Rogge, W. F.; Cass, G. R.
9 (1999) Levoglucosan, a tracer for cellulose in biomass burning and atmospheric particles. Atmos. Environ.
10 33: 173-182.
11 Spendlove, J. C. (1974) Industrial, agricultural, and municipal microbial aerosol problems. In: Current problems in
12 aerobiology: proceedings of the thirtieth general meeting of the Society for Industrial Microbiology; August
13 1973; Evanston, IL. Dev. Ind. Microbiol. 15: 20-27.
14 Standley, L. J.; Simoneit, B. R. T. (1987) Characterization of extractable plant wax, resin, and thermally matured
15 components in smoke particles from prescribed burns. Environ. Sci. Technol. 21: 163-169.
16 Susott, R. A.; Ward, D. E.; Babbitt, R. E.; Latham, D. J. (1991) The measurement of trace emissions and
17 combustion characteristics for a mass fire. In: Levine, J. S., ed. Global biomass burning: atmospheric,
18 climatic, and biospheric implications. Cambridge, MA: MIT Press; pp, 245-257.
19 Turekian, K. K. (1971) Elements, geochemical distribution of. In: Encyclopedia of science and technology, v. 4.
20 New York, NY: McGraw-Hill Book Company; pp. 627-630.
21 U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
22 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
23 U.S. Environmental Protection Agency. (2002) Health assessment document for diesel engine exhaust. Washington,
24 DC: Office of Research and Development; report no. EPA/600/8-90/057.
25 Venkataraman, C.; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons
26 and elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ. Sci.
27 Technol. 28: 555-562.
28 Vinogradov, A. P. (1959) The geochemistry of rare and dispersed chemical elements in soils. 2nd ed. New York,
29 NY: Consultants Bureau, Inc.
30 Warneck, P. (1988) Chemistry of the natural atmosphere. New York, NY: Academic Press, Inc.
31 Watson, J. G.; Chow, J. C. (1994) Clear sky visibility as a challenge for society. Annu. Rev. Energy Environ.
32 19:241-266.
33 Watson, J. G.; Fujita, E. M.; Chow, J. C.; Zielinska, B.; Richards, L. W.; Neff, W.; Dietrich, D. (1998) Northern
34 fron range air quality study final report. Reno, NV: Desert Research Institute; prepared for Colorado State
3 5 University, Cooperative Institute for Research in the Atmosphere.
36 Weingartner, E.; Burtscher, H.; Baltensperger, U. (1997a) Hygroscopic properties of carbon and diesel soot
37 particles. Atmos. Environ. 31: 2311-2327.
38 Weingartner, E.; Keller, C.; Stahel, W. A.; Burtscher, H.; Baltensperger, U. (1997b) Aerosol emission in a road
39 tunnel. Atmos. Environ. 31: 451-462.
40 Westerholm, R.; Egeback, K.-E. (1994) Exhaust emissions from light- and heavy-duty vehicles: chemical
41 composition, impact of exhaust after treatment, and fuel parameters. In: Symposium of risk assessment of
42 urban air: emissions, risk identification, and risk quantitation; May-June 1992; Stockholm, Sweden. Environ.
43 HealthPerspect. 102(suppl. 4) 13-23.
44 Westerholm, R. N.; Almen, J.; Li, H.; Rannug, J. U.; Egeback, K.-E.; Gragg, K. (1991) Chemical and biological
45 characterization of paniculate-, semivolatile-, and gas-phase-associated compounds in diluted heavy-duty
46 diesel exhausts: a comparison of three different semivolatile-phase samplers. Environ. Sci. Technol.
47 25: 332-338.
48 Wilson, T. R. S. (1975) Salinity and the major elements of sea water. In: Riley, J. P.; Skirrow, G., eds. Chemical
49 oceanography: v. 1. 2nd ed. London, United Kingdom: Academic Press Inc.; pp. 365-413.
50 Wu, J. (1979) Spray in the atmospheric surface layer: review and analysis of laboratory and oceanic results.
51 J. Geophys. Res. C: Oceans Atmos. 84: 1693-1704.
52 Yanowitz, J.; McCormick, R. L.; Grabowski, M. S. (2000) In-use emissions from heavy-duty diesel vehicles.
53 Environ. Sci. Technol. 34: 729-740.
April 2002 3D-29 DRAFT-DO NOT QUOTE OR CITE
-------
i 4. ENVIRONMENTAL EFFECTS OF
2 PARTICULATE MATTER
3
4
5 4.1 INTRODUCTION
6 Several later chapters (Chapters 5 through 8) of this document assess the latest available
7 information on determinants of human exposures to particulate matter; the dosimetry of particle
8 deposition, clearance, and retention in human respiratory tract; toxicologic evaluations of
9 pathophysiologic effects of PM and underlying mechanisms of action; and epidemiologic
10 analyses of health effects associated with human exposures to ambient PM. The human exposure
11 and health-related findings assessed in those chapters provide key elements of the scientific bases
12 to support decision making regarding review of the primary PM National Ambient Air Quality
13 Standards (PM NAAQS). This chapter, in contrast, assesses information pertinent to decision
14 making regarding secondary standards aimed at protecting against welfare effects of PM. More
15 specifically, this chapter assesses environmental effects of atmospheric PM, including PM effects
16 on vegetation and ecosystems, effects on visibility, and on man-made materials, as well as
17 relationships of ambient PM to global climate change processes.
18
19
20 4.2 IMPACTS ON VEGETATION AND ECOSYSTEMS
21 The PM NAAQS first set in 1971 were specified in terms of total suspended particulates
22 (TSP), which included both fine and coarse mode particles (the latter ranging up to 25 to 40 //m
23 in size). The 1987 revision of the PM NAAQS to PM10 standards focused attention on those
24 particles (< 10 //m mean aerometic diameter) capable of being deposited in lower (thoracic)
25 portions of the human respiratory tract. The subsequent 1997 PM NAAQS revisions retained the
26 PM10 standards and added fine particle (PM25) standards (both specified in terms of mass
27 concentrations of particles undifferentiated in terms of their specific chemical composition). The
28 effects of PM on vegetation and ecosystems as a basis for a secondary standard were not
29 considered as part of the 1997 PM NAAQS revisions. Vegetation and ecosystem effects of
30 ambient PM evaluated in this chapter are dependent not simply on PM size-related mass
April 2002 4-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentration, but rather on exposure of plants to PM components differentiated by chemical
2 composition as well.
3 This section deals with PM deposition and effects on individual plants in natural habitats
4 and terrestrial ecosystems. Except for the deposition of nitrogen and sulfur-containing
5 compounds and their effects exerted via acidic precipitation, information concerning the effects
6 of deposition of other specific substances as PM on crops is not readily available. An extensive
7 overall discussion of the effects of acidic deposition is presented in the U.S. National Acid
8 Precipitation Assessment Program (NAPAP) Biennial Report to Congress: An Integrated
9 Assessment (National Science and Technology Council, 1998). The effects of gaseous sulfur
10 oxides and nitrogen oxides on crops are discussed in detail in EPA criteria documents for those
11 substances (U.S. Environmental Protection Agency, 1982, 1993). Detailed discussion of
12 ecological effects of acidic precipitation and nitrate deposition on aquatic ecosystems can also be
13 found in the EPA Nitrogen Oxides Air Quality Criteria Document (U.S. Environmental
14 Protection Agency, 1993). Neither nitrate nor sulfate deposition on crops is discussed in this
15 chapter, as they are frequently added in fertilizers. Lead effects on crops, vegetation, and
16 ecosystems are discussed in the EPA document, Air Quality Criteria for Lead (U.S.
17 Environmental Protection Agency, 1986). Also, the effects of "certain pesticides, metal
18 compounds, chlorinated organic compounds, and nitrogen compounds" are discussed in
19 Deposition of Air Pollutants to the Great Waters, Third Report to Congress (U.S. Environmental
20 Protection Agency, 2000a).
21
22 4.2.1 Particle Deposition
23 This subsection reviews interactions between vegetation and the fine (<2.5 //m) and coarse
24 (>2.5 //m) components of airborne particulate matter (PM) that lead to deposition. Particulate
25 matter has not been defined by chemical nature, structure, or source; it has been defined mainly
26 by size fraction. While size is related to mode and magnitude of deposition to vegetated
27 landscapes and may be a useful surrogate for chemical constitution (Whitby, 1978; U.S.
28 Environmental Protection Agency, 1996a), the size classes have no specific relevance to
29 vegetation. Both fine-and coarse-mode particles may affect plants. An evaluation of particulate
30 deposition to plants and vegetated surfaces is presented because the determinants of deposition
April 2002 4-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 ultimately determine the magnitude of both contact effects and soil mediated effects of PM on
2 vegetation.
3 Particulate matter deposition to vegetation is not well understood. A recent review
4 emphasizes semivolatile organics and early work on radio nuclide deposition (Smith and Jones,
5 2000). Atmospheric deposition of particles to ecosystems takes place via both wet and dry
6 processes through three major routes: (1) precipitation scavenging in which particles are
7 deposited in rain and snow; (2) fog, cloud-water, and mist interception (i.e., "occult" deposition);
8 and (3) much slower dry deposition. Unlike gaseous dry deposition, neither the solubility of the
9 particle, nor the physiological activity of the surface are likely to be of first order importance in
10 determining particulate dry deposition velocity (Vd). Factors that contribute to surface wetness or
11 stickiness may be critical determinants of deposition efficiency. Available tabulations of
12 deposition velocity are highly variable and suspect. Recent evidence indicates that all three
13 modes of deposition (wet, occult, and dry) must be considered in determining inputs to water
14 sheds or ecosystems, because each may dominate over specific intervals of time or space.
15
16 4.2.1.1 Wet Deposition
17 Wet deposition results from the incorporation of atmospheric particles and gases into cloud
18 droplets and their subsequent precipitation as rain or snow, or from the scavenging of particles
19 and gases by raindrops or snowflakes as they fall (Lovett, 1994). Precipitation scavenging, in
20 which particles are incorporated in hydrometeors and deposited in the resulting rain and snow,
21 includes rainout (within-cloud incorporation by nucleation) and washout (below-cloud
22 scavenging by impaction). Wet deposition generally is confounded by fewer factors than dry or
23 occult deposition and has been easier to quantify. Total inputs from wet deposition to vegetative
24 canopies can be significant (Table 4-1) although not all wet deposition involves particle
25 scavenging because gaseous pollutants also dissolve in raindrops during precipitation events
26 (Lovett, 1994). This contribution is obscured during measurements because wet deposition is
27 measured simply by chemical analysis of total precipitation collected in clean, non-reactive
28 buckets. Exclusion of dry deposited material (as opposed to dissolved gaseous species) requires
29 closure or covering of the vessels except during periods of precipitation.
30 Wet deposition is not affected by surface properties as much as is dry or occult deposition
31 although leaves retain liquid and solubilized PM according to their surface properties of
April 2002 4-3 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-1. RELATIVE IMPORTANCE OF WET, DRY, PARTICIPATE, AND
TOTAL DEPOSITION TO THREE FOREST SITES3
Deposition
Total Nitrogenb
Site
Duke Forest
Gary Forest
Austin Forest
Wet
(%)
75
71
71
Dry
(%)
25
20
29
Particle
(%)
0.11
0.94
0.58
Total
(kg ha1)
9.87
5.80
6.57
Wet
(%)
64
76
83
Total Sulfur0
Dry
(%)
33
20
13
Particle
(%)
2.7
4.2
4.3
Total
(kg ha1)
17.20
7.60
7.79
"Data from Allen et al. (1994). Sampling was by triple filter pack so that fine-mode particles could be sampled
preferentially. An average particle deposition velocity of 0.9 cm s"1 was derived as in Hicks et al. (1987).
bWet nitrogen consists of NO3" and NH4+; dry nitrogen consists of vapor phase HNO3 and NO2; and paniculate
nitrogen consists of NO3".
°Wet sulfur consists of S(V2, dry sulfur consists of vapor phase SO2, and paniculate sulfur consists of pSCV2
1 wettability, exposure, and roughness. Wet deposition is largely a function of precipitation
2 amount and ambient pollutant concentrations. Any material deposited in precipitation to the
3 upper stratum of foliage will likely be intercepted by several foliar surfaces before reaching the
4 soil, because extensive vegetative canopies typically develop leaf area indices (LAI; ratio of
5 projected leaf area to ground area) much in excess of 1.
6 Landscape characteristics may also be important. Forested hillsides receive four- to
7 six-fold greater inputs of wet deposition than short vegetation in nearby valleys. This is due to a
8 variety of orographic effects (Unsworth and Wilshaw, 1989) and closer aerodynamic coupling to
9 the atmosphere of tall forest canopies than of the shorter canopies in the valleys. This leads to
10 more rapid foliar drying, which reduces the residence time but concentrates more quickly the
11 solubilized particulate materials available for foliar uptake on the cuticular surface; and
12 concentration increases the thermodynamic driving force for foliar uptake (Fowler et al., 1991;
13 Unsworth, 1984; Schonherr and Huber, 1977). Humidity and temperature conditions following
14 wet deposition strongly influence the extent of biological effects, reflecting the competing effects
15 of drying versus concentrating the solutions and influencing the rate of metabolic uptake of
16 surface solutes (Swietlik and Faust, 1984). The net consequence of these factors on direct
17 physical effects of wet deposited PM on leaves is not known.
April 2002 4-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 Rainfall introduces wet deposition and redistributes throughout the canopy previously
2 dry-deposited particulate material, particularly coarse particles which are preferentially deposited
3 in the upper foliage (Peters and Eiden, 1992). Both effects scale the likelihood of foliar contact
4 and potential direct PM effects on vegetation nearly linearly with canopy leaf area. The
5 concentrations of suspended and dissolved materials are typically highest at the onset and decline
6 with duration of individual precipitation events (Lindberg and McLaughlin, 1986; Hansen et al.,
7 1994). Sustained rainfall removes much of the accumulation of dry-deposited PM from foliar
8 surfaces, reducing direct foliar effects and combining the associated chemical burden with the
9 wet-deposited material (Lovett and Lindberg, 1984; Lovett, 1994) for transfer to the soil. Intense
10 rainfall may contribute substantial total particulate inputs to vegetated land surfaces, mostly via
11 the soil, but is less effective as a source of directly bioavailable or injurious pollutants to foliar
12 surfaces. This washing effect, combined with differential foliar uptake and foliar leaching of
13 different chemical constituents of PM, alter the composition of the rainwater that reaches the soil.
14 Low intensity precipitation events, in contrast, may be of greater significance for direct effects of
15 foliar-deposited particulate pollutants to foliar surfaces. Because of the short duration and
16 limited atmospheric cleansing, the concentration of PM in the final precipitation that remains in
17 contact with foliar surfaces may be high. Additionally, such events may hydrate some previously
18 dry-deposited particles without removing them and thereby facilitate their foliar uptake.
19 This combination of dry deposition to foliage and subsequent wet removal enhances the
20 soil pathway for PM effects, first by enhancing dry deposition relative to adjacent unvegetated
21 surfaces and then by accelerating passage along with wet deposited material of the deposited PM
22 by throughfall and stemflow to the soil where important soil-mediated, ecosystem-level
23 biogeochemical cycles of major, minor, and trace elements may be affected.
24
25 4.2.1.2 Dry Deposition
26 Dry deposition of atmospheric particles to plant and soil is a much slower process than wet
27 or occult deposition, but it acts nearly continuously and affects all exposed surfaces (Hicks,
28 1986). In dry deposition, particles at the large end of the spectrum (i.e., > 5 //m diameter) are
29 deposited mainly by gravitational sedimentation and inertial impaction. Smaller particles,
30 especially those with diameters between ~ 0.2 and 2 //m, are not readily dry-deposited and tend
31 to travel long distances in the atmosphere until their eventual deposition, most likely by
April 2002 4-5 DRAFT-DO NOT QUOTE OR CITE
-------
1 incorporation into precipitation. This long-distance transport of fine aerosols is largely
2 responsible for the regional nature of acid deposition (Lovett, 1994). A major conclusion from
3 atmospheric deposition research is the realization that dry deposition is usually a significant and,
4 in some cases, a dominant portion of total atmospheric deposition to an ecosystem (Lovett,
5 1994). Plant parts of all types, including those not currently physiologically active, along with
6 exposed soil and water surfaces, receive steady deposits of dry dusts, elemental carbon
7 encrustations, grease films, tarry acidic coatings, and heterogeneous secondary particles formed
8 from gaseous precursors (U.S. Environmental Protection Agency, 1982). The range of particle
9 sizes, the diversity of canopy surfaces, and the variety of chemical constituents in airborne PM
10 have slowed progress in both prediction and measurement of dry particulate deposition.
11 Particulate deposition is a complex, poorly characterized process controlled primarily by
12 atmospheric stability, macro- and micro-surface roughness, particle diameter, and surface
13 characteristics (Table 4-2; Hosker and Lindberg, 1982). Deposition of particles suspended
14 regionally and throughout the full depth of the planetary boundary layer (PEL) is controlled by
15 different mechanisms within the three distinct atmospheric transport zones above the surface. In
16 the lower atmosphere, fine particles are transported by turbulent eddies of mechanical and
17 convective origin. In the relatively unstirred, laminar boundary layer surrounding individual
18 surface elements, Brownian diffusion dominates. Near the surface, actual deposition and contact
19 with the surface is mediated by impaction (El-Shobokshy, 1985).
20 Deposition fluxes may be calculated from measurements, estimates, or modeled values of
21 mass concentration (C) at a specified measurement height and the total conductance or deposition
22 velocity (Vd) from this height to the surface (Eq. 4-1; Hicks et al., 1987). These modeling
23 techniques are closely allied with the micrometeorological techniques used to measure such
24 fluxes. The flux (F) may be inferred as:
25
26 F= Vd* (Cz - C0), (4-1)
27
28 where F is flux to the surface, Cz is the particle concentration at measurement height z, C0 is the
29 particle concentration at receptor sites in the canopy (usually assumed equal to 0), and Vd is the
30 overall deposition velocity. The flux is controlled by Vd and Cz.
April 2002 4-6 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-2. KEY DETERMINANTS OF DRY PARTICIPATE DEPOSITION
TO VEGETATION
Ambient concentration Proximity/strength of source
Timing/intensity of precipitation
Atmospheric conditions Wind speed/turbulence
Stability/mixing height
Temperature/humidity
Aerosol properties Chemical reactivity/solubility
Aerodynamic diameter/diffusivity/sedimentation
Biological availability
Vegetation characteristics Roughness/plant-branch spacing/flexibility
Roughness/leaf shape/pubescence
Salt/organic exudates/dew
Adapted from Lindberg and McLaughlin (1986).
1 Vertical transport of particles through the lower atmosphere to the vicinity of the vegetation
2 elements is by turbulence and sedimentation, such that:
3
4 Vd = Vt+Vs, (4-2)
5
6 in which V, (inner, left hand pathway of Figure 4-1) is a turbulent diffusion term, and Vs is a
7 sedimentation term that dominates deposition of very coarse particles (Figure 4-2) and increases
8 with particle size (Figure 4-3; dotted line). Sedimentation may be considered a pathway parallel
9 to turbulent transport (Figure 4-2), but this is an over simplification. Vs affects the concentration
10 of particles near the surface where eddy transport may occur and also governs the redeposition of
11 some fraction of the particles lost to resuspension or rebound following deposition by impaction.
12 For this reason, Vs is included (Figure 4-1) in the composite surface resistance term (RaRcpVs) as
13 well as in the parallel sedimentation term.
14 For submicron particles for which sedimentation is negligible (Hicks et al., 1987; Monteith
15 and Unsworth, 1990; Wesely, 1989), the Ohm's Law Analogy (resistance catena) analogous to
April 2002 4-7 DRAFT-DO NOT QUOTE OR CITE
-------
vd1
Particles
Atmospheric
Source
Particles
Figure 4-1. A simplified resistance catena representing the factors controlling
deposition of particles to the surface. Vegetation-specific parameters are
not explicitly considered. Modified after Hicks et al. (1987).
1
2
3
4
5
9
10
11
12
that used to describe transport of heat, momentum, or gases may be adequate, as:
Vd = Vt = [ra + rb
(4-3)
where V, is the deposition velocity due to turbulent transport of particles or other entities through
the atmosphere; rais aerodynamic resistance (inverse of conductance or velocity) associated with
the efficiency of turbulent transport above the canopy; rbis the boundary layer resistance
associated with diffusional transport through the still air layer immediately adjacent to canopy
elements; and rcis canopy resistance associated with physiological control of leaf porosity largely
stromata in the leaf surface. Significant departures from the analogy arise near the surface
(Chamberlain, 1975; Sehmel, 1980), as particles that were transported efficiently by turbulent
eddies are slowed substantially in the laminar boundary layer that reduces the efficiency of
April 2002
4-8
DRAFT-DO NOT QUOTE OR CITE
-------
1234
Coarse/Fine Ratio (|jg/|jg)
Figure 4-2. The relationship between deposition velocity of selected particulate materials
and the distribution of the material between the coarse- and fine-aerosol
fractions. Data from Foltescu et al. (1994). Ranges for Mn and Fe are from
Davidson and Wu (1989).
1 impact!on. The preservation of momentum in this zone declines with decreasing diameter;
2 however, this is offset by an increase in Brownian diffusivity with decreasing diameter
3 (Figure 4-3). Aerodynamic streamlines are parallel to the surface of each roughness element, so
4 that deposition ultimately depends on diffusion to the surface. The transition from impaction to
5 diffusion is likely blurred in the presence of leaf pubescence extending beyond the boundary
6 layer. These conflicting trends lead to a broad range over which empirical measurements of Vd
7 and particle size are relatively independent (Figure 4-3), further demonstrating the importance of
8 the quasilaminar boundary layer (Lamaud et al., 1994; Shinn, 1978).
9 The aerodynamic term (ra) decreases with increasing wind speed, turbulence, and friction
10 velocity and increases with measurement height and atmospheric stability. It describes the
April 2002
4-9
DRAFT-DO NOT QUOTE OR CITE
-------
1,000
100 —
o^
'B
c
o
10
1H
'to
8. 0.1-1
CD
Q
0.01 —
0.001
Stokes Law
Brownian Diffusion
Peters and Eiden (1992)
Little and Wiffen (1977)
0.001 0.01 0.1 1 10
Particle Diameter (|jm)
100
Figure 4-3. The relationship between particle diameter and deposition velocity for
particles. Values measured in wind tunnels by Little and Wiffen (1977) over
short grass with wind speed of 2.5 m s"1 closely approximate the theoretical
distribution determined by Peters and Eiden (1992) for a tall spruce forest.
These distributions reflect the interaction of Brownian diffusivity (descending
dashed line), which decreases with particle size and sedimentation velocity
(ascending dotted line from Stokes Law), which increases with particle size.
Intermediate-sized particles («0.1 to 1.0 (j,m) are influenced strongly by both,
and deposition is independent of size.
1 capacity of turbulent eddies to transport material, momentum, and heat between the measurement
2 height and the roughness height of the surface. Coarse particles may not be carried efficiently by
3 the high frequency eddies near the surface and may fall more rapidly than they diffuse by either
4 Brownian or turbulent process. Thus the relevance of ra breaks down as Vs increases. Indeed
5 because Vs (Eq. 4-2) is independent of a concentration gradient, the electrical analogy is a
6 theoretically flawed approximate approach (Venkatram and Pleim, 1999).
April 2002
4-10
DRAFT-DO NOT QUOTE OR CITE
-------
1 Deposition Velocity
2 Because the final stage of deposition for particles involves either impaction following
3 deceleration through a quasi-laminar boundary layer or diffusion through this boundary layer, its
4 effective depth is a critical determinant of Vd (Wiman et al., 1985; Peters and Eiden, 1992). The
5 term corresponding to the boundary layer resistance for gases (rb; equation 4-3) incorporates the
6 absence of form drag for gases. This parameter decreases with increasing turbulence and particle
7 diffusivity but is poorly characterized for gases, depending critically on canopy morphology,
8 vertical wind profiles, and gust penetration, and is of extremely limited usefulness for particles.
9 Once delivered by turbulent transport or sedimentation to the vicinity of vegetative surface
10 elements, a variety of particle size-dependent mechanisms come into play, some differing
11 substantially from those governing gaseous deposition. The concepts of rb (the still air or
12 boundary layer resistance) and rc (the canopy or surface resistance) are not generally applicable to
13 deposition of polydisperse particles. Because of the roles of momentum and bounce-off and
14 complication by reentrainment back into the airstream following deposition of a particle to the
15 surface, the factors determining the effective rb and rc for particle deposition are not as
16 independent as for gases. They are replaced in some resistance formulations (e.g., Hicks et al.,
17 1987) by the term, rcp, that combines near-surface and surface effects and by a mathematically
18 derived composite term, RaRcpVs, that combines atmospheric, surface, and sedimentation effects
19 (Figure 4-1). This latter term was insignificant for the submicron sulfate component considered
20 originally in its derivation (Hicks et al., 1987) but scales with the square of particle diameter so
21 that its general applicability to polydisperse particles is unclear. In general, transport between the
22 turbulent air column and the leaf surface through the laminar boundary layer remains difficult to
23 describe (Lindberg and McLaughlin, 1986).
24 Current estimates of regional particulate dry deposition (e.g., Edgerton et al., 1992; Brook
25 et al., 1999) infer fluxes from the product of (variable and uncertain) measured or modeled
26 particulate concentrations and (even more variable and uncertain) measured or modeled estimates
27 of dry deposition velocity parameterized for a variety of specific surfaces (e.g., Brook et al.,
28 1999). However, even for specific sites and well defined particles, uncertainties in F are largest
29 in the values of Vd, which are typically characterized by the large ranges and variances described
30 in Section 4.2.2.2 and other sources (e.g., Bytnerowicz et al., 1987a,b, Hanson and Lindberg,
31 1991, for nitrogen-containing particles; McMahon and Denison, 1979, Hicks et al., 1987, for
April 2002 4-11 DRAFT-DO NOT QUOTE OR CITE
-------
1 general treatment). The nature of the vegetative cover to which particulate deposition occurs has
2 a moderate to substantial effect on the components of Vd. The surface resistance (Hicks et al.,
3 1987) is a significant and highly site-specific component of total resistance that is difficult to
4 predict along with site, seasonal, and diurnal effects on the atmospheric components of total
5 resistance.
6 Early models of dry particulate deposition to vegetation (e.g., U.S. Environmental
7 Protection Agency, 1982; Chamberlain, 1975; Davidson and Friedlander, 1978; Garland, 1978;
8 Little and Wiffen, 1977; McMahon and Denison, 1979; Sehmel, 1980; Sehmel and Hodgson,
9 1976; and Slinn, 1977, 1978) used this paradigm (e.g., Eq. 4-3) to deal with transport to the near-
10 surface regime explicitly including conventional micrometeorological and particle size
11 considerations. Alternative modeling treatments have attempted to parameterize the geometry of
12 vegetative receptor surfaces and within-canopy micrometeorology (Wiman and Agren, 1985;
13 Peters and Eiden, 1992). Chemical reactivity, particle shape and density, rates of physiological
14 sequestration, and reentrainment by gusts of wind remain to be addressed. Modeling the
15 deposition of particles to vegetation is at a relatively early stage of development, and it is not
16 currently possible to identify a best or most generally applicable modeling approach. These
17 approaches have been further elaborated with canopy-specific choices among the available
18 models and with specific incorporation of capture efficiencies by Brook et al. (1999).
19
20 Methods of Measuring Dry Deposition
21 Methods of measuring dry deposition of particles are more restricted than for gaseous
22 species and fall into two major categories (Davidson and Wu, 1990). Surface extraction or
23 washing methods characterize the accumulation of particles on natural receptor surfaces of
24 interest or on experimental surrogate surfaces. These techniques rely on methods designed
25 specifically to remove only surface-deposited material (Lindberg and Lovett, 1985). Total
26 surface rinsate may be equated to accumulated deposition or to the difference in concentrations in
27 rinsate between exposed and control (sheltered) surfaces and may be used to refine estimates of
28 deposition (John et al., 1985; Dasch, 1987). In either case, foliar extraction techniques may
29 underestimate deposition to leaves because of uptake and translocation processes that remove
30 pollutants from the leaf surface (Taylor et al., 1988; Garten and Hanson, 1990). Foliar extraction
31 methods also cannot distinguish sources of chemicals (e.g., N) deposited as gases from those
April 2002 4-12 DRAFT-DO NOT QUOTE OR CITE
-------
1 deposited as particles (e.g., nitric acid [HNO3] or nitrate [NO3"] from nitrogen dioxide [NO2], or
2 ammonia [NH3] from ammonium [NH4+]; Bytnerowicz et al., 1987a,b; Dasch, 1987; Lindberg
3 and Lovett, 1985; Van Aalst, 1982). Despite these limitations, these methods are often used in
4 the development of in-canopy deposition models (McCartney and Aylor, 1987).
5 Deposition of pollutants by wet deposition is relatively straightforward to determine
6 through analysis of precipitation samples. Dry deposition of pollutants, on the other hand, is
7 more difficult to measure. The National Dry Deposition Network (NDDN) was established in
8 1986 to document the magnitude, spatial variability, and trends in dry deposition across the
9 United States. Currently, the network operates as a component of the Clean Air Status and
10 Trends Network (CASTNet) (Clarke et al., 1997).
11 Dry deposition is not measured directly, but is determined by a inferential approach (i.e.,
12 fluxes are calculated as the product of measured ambient concentration and a modeled deposition
13 velocity). This method is appealing and widely used because atmospheric concentrations are
14 relatively easy to measure when compared to dry deposition fluxes, and models have been
15 developed to calculate deposition velocities (Lovett, 1994). Ambient pollutant concentrations,
16 meteorological conditions, and land use data required for the inferential model are routinely
17 collected at CASTnet dry deposition sites. Chemical species include ozone, sulfate, nitrate,
18 ammonium, sulfur dioxide, and nitric acid. The temporal resolution for the ambient
19 concentration measurements and dry deposition flux calculations is hourly for ozone and weekly
20 for the other chemical substances (Clarke et al., 1997). Isotopic labeling of dry deposited PM
21 (e.g., sulfate with 35S) prior to experimental surface exposures and extractions (Garten et al.,
22 1988) can provide more precise differentiation between the deposition rates of related chemical
23 species (e.g., sulfate [SO4"2] from sulfur dioxide [SO2]).
24 At the whole-canopy level, natural surface washing by rainfall may be used to estimate dry
25 deposition of PM and gases during the preceeding dry period (Cape et al., 1992; Davidson and
26 Wu, 1990; Draaijers and Erisman, 1993; Erisman, 1993; Fahey et al., 1988; Lindberg and Lovett,
27 1992; Lovett and Lindberg, 1993; Reiners and Olson, 1984; Sievering, 1987). Collection and
28 analysis of stem flow and throughfall provides useful estimates of particulate deposition when
29 compared to directly sampled precipitation. The method is most precise for strictly PM
30 deposition when gaseous deposition is a small component of the total dry deposition and when
31 leaching or uptake of compounds of interest out of or into the foliage (i.e., N, S, base cations) is
April 2002 4-13 DRAFT-DO NOT QUOTE OR CITE
-------
1 not a significant fraction of the total deposit!onal flux (Davidson and Wu, 1990; Draaijers and
2 Erisman, 1993; Lindberg and Lovett, 1992; Lovett and Lindberg, 1993). Throughfall sampling
3 of sulfate deposition (Garten et al., 1988; Lindberg and Garten, 1988; Lindberg et al. 1990) often
4 suggests substantial foliar exchange. Other throughfall studies (e.g., Erisman, 1993; Fahey et al.,
5 1988) may lack sufficient specificity for dry particle deposition.
6 Deposition to surrogate surfaces deployed in extensive plant canopies provides a measure
7 of particle deposition to the surrounding foliage or soil surfaces. For example, a uniform
8 population of submicron particles (gold colloid, 0.8 //m) were deposited similarly to leaves of
9 Phaseolus vulgaris and to upward facing inert surfaces (Klepper and Craig, 1975). However,
10 comparison of dry deposition of particles to foliage and to inert surrogate surfaces (polycarbonate
11 Petri dishes; Lindberg and Lovett, 1985) in a deciduous forest demonstrated greater accumulation
12 on the inert surfaces; with both surfaces having accumulated particles of a similar range of sizes.
13 These persistent differences in deposition/accumulation remain to be fully characterized and
14 hinder efforts to use these surrogate techniques to provide quantitative estimates of deposition.
15 Micrometeorological methods employ an eddy covariance, eddy accumulation, or flux
16 gradient protocol in contrast to washing or extracting of receptor surfaces to quantify dry
17 deposition. These techniques require measurements of PM concentrations and of atmospheric
18 transport processes. They are currently well developed for ideal conditions of flat, homogeneous,
19 and extensive landscapes and for chemical species for which accurate and rapid sensors are
20 available. Recent progress has expanded the range of such species and extended these techniques
21 to more complex terrain (McMillen, 1988; Hicks et al., 1984; Wesely and Hicks, 1977).
22 The eddy covariance technique measures vertical fluxes of gases and fine particles directly
23 from calculations of the mean covariance between the vertical component of wind velocity and
24 pollutant concentration (Wesely et al., 1982). It is particularly limited by a requirement for
25 sensors capable of acquiring concentration data at 5-20 Hz. For the flux-gradient or profile
26 techniques, vertical fluxes are calculated from a concentration difference and an eddy exchange
27 coefficient determined at discrete heights (Erisman et al., 1988; Huebert et al., 1988). Businger
28 (1986), Baldocchi (1988), and Wesely and Hicks (1977) evaluate the benefits and pitfalls of these
29 micrometeorological flux measurements for gases. Most measurements of eddy transport of PM
30 have used chemical sensors (rather than mass or particle counting) to focus on specific PM
31 components. These techniques have not been well developed for generalized particles and may
April 2002 4-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 be less suitable for coarse PM10 transported efficiently in high frequency eddies (Gallagher et al.,
2 1988) for the same reasons that limit mathematical description of particle deposition above.
3
4 Factors Affecting Dry Deposition
5 Ambient Concentration. The ambient concentration of particles (Cz; Eq. 4-1), the
6 parameter for which there is the most data (for example, see Chapter 3, this document), is at best
7 an indicator of exposure. However, it is the amount of PM actually entering the immediate plant
8 environment that determines the biological effect. The linkage between ambient concentration
9 and delivery to vegetation is the deposition velocity (Vd), as noted above (Eq. 4-1). Cz is
10 determined by regional and local emission sources, regional circulation, and weather. It may be
11 locally sensitive to removal from the atmosphere by deposition, but the effect is generally small.
12 Current ambient PM2 5 concentrations common to non-urban ecosystems are generally well below
13 50 Aig/m. Mean annual NO3" concentrations across the eastern United States ranged from 0.2 to
14 3.9 //g/m. Summer-time mean sulfur concentrations in western states did not exceed 1 Atg/m
15 (3 Aig/m SO4"2), but mean levels in the eastern states were commonly greater than 2 //g/m,
16 especially in the southern Appalachians (Edgerton et al., 1992; Eldred and Cahill, 1994).
17 Deposition increases linearly with concentration of many materials over a broad range.
18 This allows atmospheric cleansing to take place and accounts for the greater surface impact of
19 pollutants during pollution episodes. A serious limitation of the Vd formulation used to infer
20 deposition of specific chemical species that exist in a range of particle sizes is an appropriate
21 specification of their concentration. Most sulfur emissions are readily oxidized to sulfite,
22 bisulfite, and sulfate. In the presence of atmospheric ammonia, particulate ammonium sulfate is
23 formed. However, this material is hygroscopic will increase in mass and diameter in the
24 presence of high humidity and alter its deposition behavior. Similarly, coalescence of small
25 particles into larger aggregates and adsorption of gaseous pollutants onto existing coarse particles
26 complicate the association of particle size with concentration of individual chemical species.
27 Distance and the resulting residence time in the atmosphere control the relative
28 concentrations of surface reactive materials (NO, SO2) of secondary particles that take some time
29 to form in the atmosphere and of coarse particles that exhibit high rates of deposition by
30 sedimentation near the source. These interacting processes affect the time required for formation
31 of secondary particles by gas-to-particle conversion reactions and result in a greater ratio of dry
April 2002 4-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 to wet deposition near emission sources where gaseous sulfur dioxide (gSO2) deposition
2 predominates than at greater distances where rainout of particulate SO4"2 (pSO4"2) may dominate
3 (Barrie et al., 1984) and dry deposition of pSO4"2 may be greater than of gSO2. The effect of
4 gas-to-particle conversion on dry deposition of a specific chemical species can be substantial
5 because Vd for SO2 is approximately 0.33 ± 0.17 cm s"1; whereas it is approximately
6 0.16 ± 0.08 cm s"1 for SO4"2. These phase conversions impact both Cz and the effective Vd which
7 together control dry deposit!onal fluxes (Eq. 4-1). The neutralization of acidic gaseous and
8 particulate species by alkaline coarse particles has been described in arid regions, but it may be
9 more prevalent in urban New York, where coarse particles are observed to be neutral with
10 alkaline cations approximately balancing gaseous acidic species (Lovett et al., 2000). The
11 deposition of the acidic materials in the urban environment is likely enhanced by incorporation
12 into these previously formed coarse particles.
13 Similarly, the ratio of coarse to fine particle concentrations determines the effective Vd for
14 chemically speciated particles (Figure 4-2). This reflects the size-dependent deposition processes
15 that govern delivery of PM to receptor surfaces (Fig. 4-3). For example, SO4"2 was found
16 predominantly on fine submicron particles; whereas potassium ion (K+), calcium (Ca+2), and
17 nitrate (NO3") were associated most often with coarse particles larger than 2 //m (Lindberg and
18 Lovett, 1985). However, concentrations of particulate S and K+ within a coniferous canopy were
19 strongly correlated (Wiman and Lannefors, 1985), suggesting a primary source of coarse-mode
20 sulfur particles. Furthermore, NO3" and SO4"2-containing fine particles readily coalesce with
21 coarse particles derived from sea spray or primary geologic material (Wu and Okada, 1994;
22 Milford and Davidson, 1987). Gaseous N and S species may undergo gas to particle conversion
23 directly onto such preexisting coarse particles. As a result, marine and continental particle size
24 spectra for both N and S differ substantially, with a peak in the coarse mode generally apparent
25 near marine sources (Milford and Davidson, 1987). The issue for NO3" is further confounded by
26 uncertain discrimination between gaseous and particulate species in current sampling methods.
27 The substantial effect of particle size on Vd (Figure 4-3) implies a need for size resolution as well
28 as chemically speciated ambient particulate concentrations even within the PM10 fraction.
29
30 Particle Effects on Vd. Particle size is a key determinant of Vd as noted above; but,
31 unfortunately, the size spectra may be quite complex. The particles in the study of Lindberg and
April 2002 4-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lovett (1985) at Walker Branch Watershed had median diameters ranging from 3 to 5 //m; but
2 approximately 25% of the particles had diameters less than 1 //m (0.2 to 0.3 //m), and 5 to 20%
3 of the particles were much larger aggregates. The aggregated particles are significant in that
4 chemically they reflect their fine particle origins, but physically they behave like large particles
5 and deposit by sedimentation. Direct observations with SEM demonstrate that particle
6 morphology can be highly variable. Many submicron particles can be observed on trichomes
7 (leaf hairs), although most particles are in the 5 to 50 //m diameter range. Large aggregated
8 particles in excess of 100 //m also are seen, with carbonaceous aggregate particles being
9 especially common (Smith, 1990a). Trichomes are especially efficient particle receptors;
10 however, trichomes are reduced in size by "weathering" and occasionally are completely broken
11 off during the growing season.
12 In the size range around 0.1 to 1.0 //m, where Vd is relatively independent of particle
13 diameter (Fig. 4-3), deposition is controlled by macroscopic roughness properties of the surface
14 and by the stability and turbulence of the atmospheric surface layer. The resistance catena
15 (Figure 4-1) is less useful in this size range and, in some treatments, has been abandoned entirely
16 (e.g., Erisman et al., 1994; Eq. 4-4). Impaction and interception dominate over diffusion, and the
17 Vd is considerably (up to two orders of magnitude; Figure 4-3) lower than for particles either
18 smaller or larger (Shinn, 1978). The deposition velocity may be parameterized in this size range
19 as a function of friction velocity,
20
21 Vd = (a/b)u*, (4-4)
22
23 where a depends on atmospheric stability and b depends on surface roughness (Wesely et al.,
24 1985; Erisman et al., 1994). Similar formulations have been presented in terms of turbulence
25 (standard deviation of wind direction) and wind speed (e.g., Wesely et al., 1983), both
26 determinants of u*.
27 Deposition of particles between 1 and 10 //m diameter, including the coarse mode of PM10,
28 is strongly dependent on particle size (Shinn, 1978). Larger particles within this size range are
29 collected more efficiently at typical wind speeds than are smaller particles (Clough, 1975),
30 suggesting the importance of impaction. Impaction is related to wind speed, the square of
31 particle diameter, and the inverse of receptor diameter as a depositing particle fails to follow the
April 2002 4-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 streamlines of the air in which it is suspended around the receptor. When particle trajectory
2 favors a collision, increasing wind speed or ratio of particle size to receptor cross section makes
3 collision nearly certain; and, as these parameters become very small, the probability of collision
4 becomes negligible. However, the shape parameter for the more common range of situations
5 between these extremes remains poorly characterized (Peters and Eiden, 1992; Wiman and
6 Agren, 1985).
7 As particle size increases above 1 //m, deposition is governed increasingly by
8 sedimentation (Figure 4-3) with a correspondingly declining influence of turbulence and
9 impaction. Particles between approximately 10 and 24 //m (Gallagher et al., 1988) are both
10 small enough to be transported efficiently by turbulent eddies to the surface and large enough to
11 impact with sufficient momentum to overcome boundary layer effects. These particles deposit
12 highly efficiently and relatively independently of particle size.
13 Deposition of the largest suspended particles (e.g., >50 //m) is governed, through
14 sedimentation and the corresponding terminal settling velocity (Vs), almost entirely by size.
15 These particles are not transported efficiently by small-scale eddies near the surface.
16 Theoretically based models for predicting particle deposition velocities have been
17 published by Bache (1979a,b), Davidson et al. (1982), Noll and Fang (1989), Slinn (1982), and
18 Wiman (1985). These models deal primarily with low canopies or individual elements of canopy
19 surfaces. Wiman and Agren (1985) have developed an aerosol deposition model that specifically
20 treats the problem of particle deposition to forests where turbulence plays a particularly
21 important role, especially at roughness transitions such as forest edges. They found that
22 deposition of supermicron particles is controlled by complex interactions among particle size and
23 concentration, forest structure, and aerodynamics; whereas deposition of fine particles
24 (submicron) is controlled by particle concentration and forest structure.
25 At the present time, empirical measurements of Vd for fine particles under wind tunnel and
26 field conditions are often several-fold greater than predicted by available theory (Unsworth and
27 Wilshaw, 1989). A large number of transport phenomena, including streamlining of foliar
28 obstacles, turbulence structure near surfaces, and various phoretic transport mechanisms remain
29 poorly parameterized in current models. The discrepancy between measured and predicted
30 values of Vd may reflect such model limitations or experimental limitations in specification of
31 the effective size and number of receptor obstacles, as suggested by Slinn (1982).
April 2002 4-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 Available reviews (Davidson and Wu, 1990; McMahon and Denison, 1979; Nicholson,
2 1988; Sehmel, 1980; Slinn, 1982; U.S. Environmental Protection Agency, 1982, 1996a) suggest
3 the following generalizations: (1) particles larger than 10 //m exhibit a variable Vd between «0.5
4 and 1.1 cm s"1 depending on friction velocities; whereas a minimum particle Vd of 0.03 cm s"1
5 exists for particles in the size range 0.1 to 1.0 //m; (2) the Vd of particles is approximately a
6 linear function of friction velocity; and (3) deposition of particles from the atmosphere to a forest
7 canopy is from 2 to 16 times greater than deposition in adjacent open terrain (i.e., grasslands or
8 other vegetation of low stature).
9
10 Leaf Surface Effects on Vd. The term rc (Equation 4-3) reflects the chemical, physical, or
11 physiological characteristics of the surface that govern its ability to capture, denature, or
12 otherwise remove paniculate material from the atmospheric surface layer. For gases, relevant
13 surface properties involve the physiological state of the vegetation, including stomatal opening
14 and mesophyll antioxidant activity, and the chemical reactivity of the exposed surface with the
15 specific gas. For particles, relevant surface properties involve stickiness, microscale roughness,
16 and cross-sectional area which determine the probability of impaction and bounce (e.g., Shinn,
17 1978). The chemical composition of the PM is not usually a primary determinant of deposition
18 velocity. At the microscopic scale where Van der Waals forces may determine particle bounce
19 and reentrainment, the chemical properties of both surface and particle may be significant but
20 remain poorly understood.
21 Stickiness may itself depend on previous deposition of deliquescent particles that prolong
22 leaf wetness, on the wettability of foliar surfaces, and on the presence of sticky residues such as
23 honeydew deposited by aphids. These factors increase deposition by decreasing bounce-off,
24 wind reentrainment, and, to some extent, wash-off by precipitation.
25 The distribution of particles on and the efficiency of deposition to vegetation also varies
26 based on leaf shape and plant part. Particles are more prevalent on the adaxial (upper [facing the
27 twig]) surface than on the abaxial (lower [away from the twig]) surface. Peripheral leaf areas
28 tend to be the cleanest with most particles accumulating in the midvein, central portion of leaves.
29 The rough area surrounding the stomatal pores was not found to be a preferential site for
30 particle deposition or retention (Smith and Staskawicz, 1977). Most particles were located near
31 veins with smaller particles localized on the trichomes. The greatest particulate loading on
April 2002 4-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 dicotyledonous leaves is frequently on the adaxial surface at the base of the blade, just above the
2 petiole junction. It is probable that precipitation washing plays an important role in this
3 distribution pattern. Lead particles accumulated to a larger extent on older than younger needles
4 and twigs of white pine, indicating that wind and rain were insufficient to fully wash the foliage.
5 Fungal mycelia (derived from windborne spores) were frequently observed in intimate contact
6 with other particles on other leaves (Smith and Staskawicz, 1977), which may reflect shelter by
7 the particles minimizing reentrainment of the spores, mycelia development near sources of
8 soluble nutrients provided by the particles, or codeposition. This pattern is significant and could
9 yield further insight into deposition mechanisms.
10 Leaves with complex shapes collect more particles than those with regular shapes. Conifer
11 needles are more effective than broad leaves in accumulating particles. The edge to area ratio
12 (Woodcock, 1953) is also a key determinant of salt deposition to individual artificial leaves.
13 A strong negative correlation was observed under wind tunnel conditions between the area of
14 individual leaves and deposition of coarse particles (Little, 1977). Small twigs and branches
15 were more effective particle collectors than were large branches and trunks of trees (Smith,
16 1984). Lead particles accumulated 20-fold more on woody stems than on leaves of white pine
17 (Pinus strobus), even though leaves displayed a 10-fold greater total area (Heichel and Hankin,
18 1976). Deposition is heaviest at tips of individual leaves.
19 Rough, pubescent broadleaf discs collected coarse (5.0-//m) particles up to sevenfold more
20 efficiently than glabrous leaf discs (Little, 1977). Laminae, petioles, and stems, all differed in
21 collection efficiency. Pubescent leaves of sunflower (Helianthus annuus) collected coarse
22 particles nearly an order of magnitude more efficiently than the glabrous leaves of tulip poplar
23 (Liriodendron tulipiferd) under wind tunnel conditions (Wedding et al., 1975). Rough pubescent
24 leaves of nettle (Urtica dioicd) were more effective in capturing coarse PM10 than were the
25 densely tomentose leaves of poplar (Populus alba) or smooth leaves of beech (Fagus sylvaticd).
26 Conifer needles are more efficient than broad leaves in collecting particles by impaction.
27 This reflects the small cross section of the needles relative to larger leaf laminae of broadleaves
28 and the greater penetration of wind into conifer than broadleaf canopies (below). Conifers were
29 more effective in removing coarse (-20 //m) particles of ragweed pollen from the atmosphere
30 than were broadleaf trees (Steubing and Klee, 1970) and in intercepting the even coarser particles
31 of rain (Smith, 1984). Conifers are also more effective in retaining and accumulating particles
April 2002 4-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 against reentrainment by wind and removal by rain, particularly on foliar surfaces where they are
2 likely to be most biologically active. Seedlings of white pine (Pinus strobus) and oak (Quercus
3 rubrd) each initially retained between a quarter (pine) and a third (oak) of very coarse (88 to
4 175 //m) 134Cs-tagged quartz particles applied under field conditions (Witherspoon and Taylor,
5 1971). After 1 h, the pine retained over 20% of the 134Cs-tagged particles; whereas the oak
6 retained only approximately 3%. Long-term retention of the particles was concentrated at the
7 base of the fascicles in pine and near the surface roughness caused by the vascular system on
8 leaves of oak. The sheltered locations available in the conifer foliage contribute substantially to
9 greater retention of particles. For similar reasons, grasses also are efficient particle collectors
10 (Smith and Staskawicz, 1977) with long-term retention mostly in the ligule and leaf sheath.
11 Wind tunnel studies also demonstrated equivalent deposition properties of 3.36-//m
12 particles of dense lead chloride and 6.11-^m particles of less dense uranine dye. These particles
13 were shown to be aerodynamically equivalent, substantiating the use of aerodynamic diameter as
14 a classification parameter for particle deposition.
15
16 Canopy Surface Effects on Vd. In general, surface roughness contributes to greater
17 particulate deposition. As a result, Vd is typically greater for a forest than for a field or
18 nonwoody wetland and greater for a field than for a water surface. The contrasting transport
19 properties and deposition velocities of different size particles lead to predictable patterns of
20 deposition. For coarse particles, the upwind leading edges of forests, hedge rows, and individual
21 plants, as well as of individual leaves, are primary sites of deposition. Impaction at high wind
22 speed and the sedimentation that follows the reduction in wind speed and carrying capacity of the
23 air in these areas lead to preferential deposition of larger particles.
24 Air movement is slowed in proximity to vegetated surfaces. Resulting log profiles of wind
25 and pollutant concentrations in the near-surface turbulent boundary layer above canopies reflect
26 surface characteristics of roughness length, friction velocity, and displacement height. Plasticity,
27 streamlining, and oscillations of foliar elements also alter the aerodynamic roughness and the
28 level of within-canopy turbulence. Canopies of uneven age or with a diversity of species are
29 typically aerodynamically rougher and receive larger inputs of pollutants than do smooth, low, or
o
30 monoculture vegetation (Garner et al., 1989; Wiman and Agren, 1985). Canopies on slopes
31 facing the prevailing winds and individual plants on the windward edges of discontinuities in
April 2002 4-21 DRAFT-DO NOT QUOTE OR CITE
-------
1 vegetative cover over which roughness increases receive larger inputs of pollutants than more
2 sheltered, interior canopy regions. For example, some 80% of coarse particulate sea salt was
3 deposited on the upwind edge of a hedgerow (Edwards and Claxton, 1964), and the concentration
4 of ragweed (Ambrosia spp.) pollen was reduced by 80% within 100 m of the leading edge of a
5 forest (Neuberger et al., 1967).
6 Beier et al. (1992) and Beier (1991) discussed two methods for estimating the dry
7 deposition of base cations to forest edges: (1) a difference method between measured
8 precipitation and throughfall concentrations of base cations, and (2) a calculation method based
9 on known ratios of Na+ deposition in wet and dry forms (Ulrich, 1983). A combination of these
10 two approaches produced the best estimates of SO4"2, Ca+2, Mg+2, and K+ particle deposition.
11 Using these methods, particulate SO4"2 (Beier, 1991) and particulate Ca+2, Mg+2, and K+ (Beier
12 et al., 1992) were found to decrease by an order of magnitude from the forest edge to the forest
13 interior. A number of authors also have shown that particle deposition is elevated at forest edges
14 when compared to a uniform forest canopy (Draaijers et al., 1988; Grennfelt, 1987; Lindberg and
15 Owens, 1993), and Draaijers et al. (1992) reported that differences are likely to exist between
16 forest types because of variable canopy structure. Draaijers et al. (1988) further emphasized that
17 enhanced particle deposition at or near forest edges is strongly dependent on the velocity and
18 wind direction during observations.
19 The factors leading to horizontal gradients are confounded by time- and distance-related
20 sedimentation, geologic dust (mostly around 7 //m aerodynamic diameter) being collected on
21 stems of wild oats (Avena spp.; Davidson and Friedlander, 1978) and on eastern white pine
22 (Pinus strobus; Heichel andHankin, 1972; Smith 1973) downwind of roadways. Rapid
23 sedimentation of coarse crustal particles suggests that potential direct effects may be restricted to
24 roadway margins, forest edges, and, because of the density of unpaved roads in agricultural areas,
25 crop plants.
26 Simulated deposition to an ecologically complex, mixed canopy was considerably higher
27 than to a pure spruce stand in which most of the leaf area was concentrated in regions of low
28 wind speed. Limitations to the application of these models to predict deposition over large
29 regions include a limited understanding both of the nature of microscopic particle-surface
30 interactions and of the effects of complex terrain and species composition on macroscopic
31 transport processes.
April 2002 4-22 DRAFT-DO NOT QUOTE OR CITE
-------
1 Macroscopic turbulent transport processes, related to ra, at successive layers through the
2 canopy can be separated from microscopic processes, related to rb and rc (or rcp), at each
3 deposition surface (e.g., Peters and Eiden, 1992; Wiman and Lannefors, 1985). The macroscopic
4 approach deals with deposition as the product of a turbulent diffusion coefficient and a
5 concentration (Cz) at each canopy layer, both of which vary with particle size and with height (Z)
6 in the canopy. The microscale parameters involve those factors that determine absorption of a
7 particle at each surface as captured imperfectly by rc. Shelter effects caused by the crowding of
8 foliar elements within the canopy can be ignored if the wind speed within each canopy layer is
9 specified. This approach requires knowledge of the vertical distribution of particle concentration
10 and foliage density in the canopy airspace along with profiles of wind speed or turbulence.
11 Once introduced into a forest canopy, elements associated with course particles tend to
12 decrease markedly with canopy depth; whereas elements associated with fine particles do not
13 (Lovett and Lindberg, 1992).
14 Trace elements and alkaline earth elements are enriched below the canopies of both
15 southern (Lindberg et al., 1986) and northern (Eaton et al., 1973) hardwood forests. Vertical
16 gradients in concentration of coarse particles and of elements associated with coarse particles
17 were observed in a mixed conifer/birch forest canopy (Wiman and Lannefors, 1985; Wiman
18 et al., 1985) and in a mixed oak forest (e.g., Ca+2, Figure 4-4A; Lovett and Lindberg, 1992). The
19 highly reactive gas HNO3 also exhibited a vertical gradient, but with a steep decline at the top of
20 the canopy (Figure 4-4B). Lovett and Lindberg (1992) studied concentration profiles of various
21 gases and particles within an closed canopy forest and concluded that coarse particle
22 concentrations associated with elements like Ca+2 would decrease markedly with depth in the
23 canopy, but they found only minor reductions with depth in the concentrations of fine aerosols
24 containing SO4"2, NH4+, and FT. These data suggest that all foliar surfaces within a forest canopy
25 are not exposed equally to particle deposition: upper canopy foliage would receive maximum
26 exposure to coarse and fine particles, but foliage within the canopy would receive primarily fine
27 aerosol exposures. Fine-mode particles (e.g., sulfate [SO4"2], Figure 4-4C) and unreactive gases
28 typically do not exhibit such vertical profiles, suggesting that uptake is smaller in magnitude and
29 more evenly distributed throughout the canopy. In multilayer canopies, simultaneous
30 reentrainment and deposition may effectively uncouple deposition from local concentration.
31
April 2002 4-23 DRAFT-DO NOT QUOTE OR CITE
-------
g>
'o>
X
25 -
20 -
15 -
10 -
5 -
0 -
A = 34%
A = 43%
0.6 0.7 0.8 0.9
D)
'0
I
25 -
20 -
15 -
10 -
5 -
0 -
A = 6%
D
A = 0%
9.5
10.0 0.08 0.10
Concentration (|jg/m3)
0.12 0.14
Figure 4-4. Vertical stratification of diverse, chemically speciated particles in a mixed oak
forest. Nitric acid vapor, as a highly reactive, nonparticulate species, is shown
for comparison. The horizontal dotted line indicates canopy height, and A
indicates the percent depletion from above to below the canopy. Data from
Table 2-4 of Lovett and Lindberg (1992).
1 Polydisperse size distributions of many chemical species effectively prevent the use of a single
2 estimate of Vd for any element if highly accurate results are required.
3 Although gradients (Figures 4-4 and 4-5) may be related to local Vd within the canopy
4 (Bennett and Hill, 1975), the absence of a gradient may reflect either low rates of deposition or
5 very high rates relative to turbulent replenishment from above the canopy (Tanner, 1981).
6 Below- or within-canopy emissions may confound interpretation of vertical gradients. Linear
7 gradients of the gaseous pollutants hydrogen fluoride (HF) and ozone (O3) reflected large uptake
8 rates; whereas small gradients in NO suggested little uptake by foliage (Bennett and Hill, 1973,
April 2002
4-24
DRAFT-DO NOT QUOTE OR CITE
-------
c
o
'*—>
03
1.0-
0.8-
0) CD
O .^
O TO
^ ^
CD Q. U.b—E
?°
t 03
03 O
D- Q)
£S
O -°
c ro
03 xo
O ^
0.4-
0.2-
GO
0-
wtlllllllllllll111"" " IIIIIIIIIIIIIIH/^
/* \
0.01 1
Particle Diameter (|jm)
100
Figure 4-5. The relationship between particle size and concentration below a spruce
canopy with wind velocity at a height of 16.8 m equaling 5 m s"1. Modified
from Peters and Eiden (1992).
1 1975). However, soil efflux of NO could have complicated the latter interpretation. The lack of
2 a vertical gradient and a peak near the top of the active canopy in particulate K+ (Figure 4-4D)
3 was interpreted as evidence for a biogenic source within the deciduous forest canopy with
4 moderate rates of deposition (Lindberg et al., 1986; Lovett and Lindberg, 1992).
5 The size dependence of this vertical stratification of parti culate concentration (see
6 Figure 4-5) is reflected in current simulation models (Wiman et al., 1985; Peters and Eiden,
o
7 1992). The model of Wiman and Agren (1985) predicts a uniform vertical distribution of
8 fine-mode particles and a pronounced vertical gradient of coarse-mode particles which is in
9 agreement with observations (Lovett and Lindberg, 1992).
10 Simulation of the horizontal deposition patterns at the windward edge of a spruce forest
11 downwind of an open field with the canopy between 1 and 25 m above the ground indicated that
12 deposition was maximal at the forest edge where wind speed and impaction were greatest.
13 Simulation of the vertical deposition pattern was more complex. Deposition was not greatest at
April 2002
4-25
DRAFT-DO NOT QUOTE OR CITE
-------
1 the top of the canopy where wind speed was highest, but at z = 0.75 h where the balance between
2 leaf area (obstacles for impaction) and wind speed (momentum for impaction) was optimal,
3 although neither parameter alone was maximal. Simulated deposition in this spruce forest
4 increased considerably with increasing LAI at the forest edge, where wind speed was insensitive
5 to LAI but the number of obstacles increased. Inside the forest, where both wind speed and
6 impaction increasingly were attenuated by increasing LAI, deposition increased only marginally
7 in spite of the increase in obstacle frequency.
8 To scale surface-specific measurements of particle deposition to forest or crop canopies,
9 conversions of the following type have been suggested:
10
Vd.canopy = Vd surface * scaling factor, (4-5)
11
12 with empirical scaling factors proposed by Lindberg et al. (1988).
13 To appropriately scale surface-specific measurements of particle deposition to landscapes,
14 one must consider the complexity of grassland, crop, and forest canopies in order to avoid serious
15 over- or under-estimates of particle deposition. Individual species exposed to similar ambient
16 concentrations may receive a range of particulate loading that is more closely related to foliar
17 damage than the ambient concentration (Vora and Bhatnagar, 1987).
18 Both uptake and release of specific constituents of PM may co-occur within a single canopy
19 (e.g., K+; Lovett and Lindberg, 1992). The leaf cuticular surface is a region of dynamic exchange
20 processes through leaching and uptake. Exchange occurs with epiphytic microorganisms and
21 bark and through solubilization and erosion of previously deposited PM. Vegetation emits a
22 variety of parti culate and parti culate precursor materials. Terpenes and isoprenoids predominate
23 and, on oxidation, become condensation nuclei for heterogeneous particle formation. Salts and
24 exudates on leaves and other plant parts continually are abraded and suspended as particles, as
25 are plant constituents from living and dead foliage (Rogge et al., 1993a). Soil minerals,
26 including radioactive strontium, nutrient cations and anions, and trace metals are transferred to
27 the active upper foliage and then to the atmosphere in this way. Although not representing a net
28 addition to an ecosystem, particle release from vegetation is a mechanism for redistributing
29 chemical pollutants derived from the soil or prior deposition within a canopy, potentially
30 enhancing direct effects and confounding estimates of Vd.
April 2002 4-26 DRAFT-DO NOT QUOTE OR CITE
-------
1 Range of Deposition Velocity
2 As noted in the previous criteria document (U.S. Environmental Protection Agency, 1982)
3 and in McMahon and Denison (1979), estimates of Vd for PM10 particles to vegetation are
4 variable and suggest a minimum between 0.1 and 1.0 //m as predicted from first principles
5 (Monteith and Unsworth, 1990; Sehmel, 1980). Determinations in wind tunnels with passive
6 collectors and micrometeorological methods tend to converge in this range. The range of Vd for
7 sulfate from passive collectors was found to be from 0.147 to 0.356 cm s"1; and, from eddy
8 covariance techniques, a mean Vd of 0.27 cm s"1 was observed (Dolske and Gatz, 1984).
9 Micrometeorological techniques over grass (Wesely et al., 1985); indirect, inert collector
10 techniques within an oak forest (Lindberg and Lovett, 1985); and many other empirical
11 determinations (e.g. McMahon and Denison, 1979; Table 4-3) generally support this range. Over
12 aerodynamically smooth snow (Duan et al., 1988; Table 4-4), measurements of Vd were an order
13 of magnitude smaller. Very coarse particles, often non-size-specified primary geologic material,
14 frequently exhibit Vd greater than 1.0 cm s"1 (e.g., Clough, 1975). The increase in Vd with
15 decreasing size below 0.1 //m is probably hidden in most empirical determinations of Vd, because
16 the total mass in this fraction is very small despite the large number of individual particles.
17 Table 4-5 shows published estimates of Vd with variability estimates for fine particles of
18 specified aerodynamic diameters dominated by a range of chemical species.
19 Ibrahim et al. (1983) evaluated the deposition of ammonium sulfate particles to a range of
20 surfaces and found that particles having a mean diameter of 0.7 //m had deposition velocities
21 ranging from 0.039 to 0.096 cm s"1. Larger particles (having mean diameters of 7 //m) had
22 greater deposition velocities (between 0.096 and 0.16 cm s"1). The authors further concluded that
23 the hygroscopic nature of the sulphate particle could increase its size and enhance deposition near
24 sources of water, e.g., snow. Using eddy correlation approaches, Hicks et al. (1989) found a
25 mean daily Vd for sulfur-containing PM to be 0.6 cm s"1. However, they suggested that the Vd
26 value could be as high as 1 cm s"1 during the day and near zero at night.
27 Lindberg et al. 1990 found a wide discrepancy between deposition velocities for NO3"
28 between study sites in Oak Ridge, TN (~2 cm s"1) and Gottingen, Germany (-0.4 cm s"1). They
29 suggest that the increased Vd at Oak Ridge could be explained by the primary occurrence of NO3"
30 in coarse particles that exhibit greater Vd than fine particles (Davidson et al., 1982). Large values
April 2002 4-27 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-3. REPORTED MEAN DEPOSITION VELOCITIES (Vd) FOR SULFATE,
CHLORINE, NITRATE, AND AMMONIUM AND ION-CONTAINING PARTICLES
Chemical Species/ Surface
S042
Inert plates
Inert plates
Inert plates
Inert bucket
Foliage
Chaparral
Grass canopy
Grass canopy
Pine foliage
Plant canopies
Grass canopy
Cl
Inert bucket
Inert bucket
Beech canopy
Spruce canopy
NO3
Inert plates
Ceanothus
Glycine max
Ligustmm
Quercus
Quercus
summer
winter
Quercus
summer
winter
Pinus
Pasture
Ulmus
NH4+
Calluna/Molina
Ceanothus
Kalmia
Pinus
V^cms-'r
0.13
-0.14
0.14
0.51
0.29
0.15
0.10
0.07
0.07
0.50
0.22
3.1
5
1
1.9
0.4-2
0.4
0.24
0.1-0.5
0.7-1.1
0.55
0.7
0.3
0.1
0.5-1.3
0.7-0.8
1.1
0.18
0.4
0.03-0.14
0.01 -0.06
Method
Extraction
Extraction
Extraction
Extraction
Extraction
Extraction
Gradient
Gradient
Extraction
Gradient
Eddy covariance
Extraction
Extraction
Throughfall
Throughfall
Extraction
Extraction
Extraction
Extraction
Extraction
Throughfall
Throughfall
—
—
Extraction
Gradient
Extraction
Gradient
Extraction
Extraction
Extraction
Reference
Lindberg and Lovett (1985)
Lindbergetal. (1990)
Davidson and Wu (1990)a
Davidson and Wu (1990)a
Davidson and Wu (1990)a
Bytnerowicz et al. (1987a)
Allen etal. (1991)
Nicholson and Davies (1987)
Wiman(1981)
Davidson and Wu (1990)a
Weseley etal. (1985)
Dasch and Cadle ( 1985)
Dasch and Cadle ( 1986)
Hofken etal. (1983)
Hofken etal. (1983)
Lindbergetal. (1990)
Bytnerowicz et al. (1987a)
Dolske (1988)
John etal. (1985)
Dasch (1987)
Lovett and Lindberg (1984)
Lovett and Lindberg (1984)
Lovett and Lindberg (1986)
Lovett and Lindberg (1986)
Dasch (1987)
Huebert etal. (1988)
Dasch (1987)
Duyzer etal. (1987)
Bytnerowicz et al. (1987a)
Tjepkema et al. (1981)
Dasch (1987)
"These data represent the mean of data by measurement technique as reported in the cited reference. The reader is referred to the referenced
articles for information on the specific cations contributing to the means.
April 2002
4-28
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-4. REPRESENTATIVE EMPIRICAL MEASUREMENTS OF DEPOSITION
VELOCITY (Vd) FOR PARTICIPATE DEPOSITION
vd
x ± SE (cm s'1)
0.034 ±0.014
0.021 ±0.005
0.1 ±0.03
0.22 ± 0.06
0.13 ±0.02
0.75 ± 0.24
l.liO.l
0.9
2.5
9.4
Particle Size
C"m)
0.15-0.30
0.5-1.0
0-2
0.1-2.0
(so42-)
(K+)
(Ca+2)
2.75
5.0
8.5
Method
Eddy covariance with optical counter, flat
snow surface
Profile, fine SO42", short grass
Eddy covariance with flame photometer plus
denuder, 40-cm grass, fine SO42"
Inert surface collectors (petri dish) in oak
forest
Wind tunnel to pine shoots; polystyrene
beads; within- "canopy" wind speed, 2.5 m s"1
Reference
Duanetal. (1988)
Allen etal. (1991)
Weselyetal. (1985)
Lindberg and Lovett
(1985)
Chamberlain and Little
(1981)
TABLE 4-5. REPORTED MEAN DEPOSITION VELOCITIES FOR POTASSIUM,
SODIUM, CALCIUM, AND MAGNESIUM BASE CATION CONTAINING
PARTICLES
Chemical Species/Surface
K+
Inert plates
Inert bucket
Na+
Inert bucket
Inert plate
Ca+2
Inert plates
Inert plates
Inert bucket
Mg+2
Inert bucket
VaCcms-1)
0.75
0.51 -2.4
1.7-2.9
0.8-8.2
1.1
»2
1.7-3.2
1.1 -2.7
Method
Extraction
Extraction
Extraction
Extraction
Extraction
Extraction
Extraction
Extraction
Reference
Lindberg and Lovett (1985)
Dasch and Cadle (1985)
Dasch and Cadle (1985)
Lindberg and Lovett (1985)
Lindberg etal. (1990)
McDonald etal. (1982)
Dasch and Cadle (1985)
April 2002
4-29
DRAFT-DO NOT QUOTE OR CITE
-------
1 of Vd for base-cation-containing particles (>1 cm s"1) suggest their occurrence in coarse particles
2 (Lindberg and Lovett, 1985).
3 The several attempts to estimate Vd for SO4"2, NO3", and NH4+ with the throughfall mass
4 balance approach (Davidson and Wu, 1990; Gravenhorst et al., 1983; Hofken and Gravenhorst,
5 1982) have produced higher Vd values that are considered suspect. They have not been included
6 in Tables 4-3 and 4-5. Overestimates of Vd for SO4"2 and NO3" particles derived from throughfall
7 mass balance approaches may be the result of gaseous SO2 and HNO3 gaseous deposition to
8 foliar surfaces (Lindberg and Lovett, 1985). A similar contribution of NH3 deposition may lead
9 to erroneously high Vd values for NH4+ when the throughfall method is attempted in areas of high
10 NH3 concentrations. Dolske's (1988) reported Vd values for NO3" deposition to soybean ranged
11 from 0.4 to 31 with a mean of 0.24 cm s"1. However, because Dolske's leaf extraction
12 measurements included a component of HNO3 vapor, the Vd values may represent more than
13 deposition caused by aerosol nitrate alone.
14
15 4.2.1.3 Occult Deposition
16 Gaseous pollutant species may dissolve in the suspended water droplets of fog and clouds.
17 The stability of the atmosphere and persistence of the droplets often allow a condition of
18 gas/liquid phase equilibrium to develop. This allows knowledge of air mass history or ambient
19 concentrations of specific pollutants to be used to estimate fog or cloud water concentrations.
20 Further estimates of the deposition velocity of the polluted droplets allows calculation of
21 deposit!onal fluxes. Unfortunately, interception of fog or cloud droplets by plant parts or other
22 receptor surfaces remains difficult both to predict and to measure. Fog formation influences the
23 total atmospheric burden and deposition of paniculate matter (Pandis and Seinfeld, 1989) by
24 accreting and removing particles from the air, by facilitating particle growth through aqueous
25 oxidation reactions, and by enhancing deposition as noted. Aqueous condensation may occur
26 onto preexisting fine particles, and such particles may coalesce or dissolve in fog or cloud
27 droplets. Material transported in fog and cloud water and intercepted by vegetation escapes
28 detection by measurement techniques designed to quantify either dry or wet deposition; hence it
29 is hidden (i.e., "occult") from the traditional measurements.
30 Low elevation radiation fog has different formation and deposition characteristics from
31 high elevation cloud or coastal fog water droplets. A one dimensional deposition model has
April 2002 4-30 DRAFT-DO NOT QUOTE OR CITE
-------
1 recently been described for a radiation fog episode (Von Glasow and Bott, 1999). A substantially
2 greater concentration of key polluting species (eg. NO3", SO4"2, organics) may be observed in
3 smaller than in larger droplets in fog (Collett et al., 1999). Acidity differences exceeding 1 pH
4 unit were also observed in the San Joaquin Valley winter radiation fog with smaller particles
5 being more acidic. This has implications for aqueous phase oxidation of sulfur and nitrogen
6 compounds, in particular, while sulfur oxidation by ozone (the dominant reaction in this
7 environment even during winter) is well known in typically acidic fog droplets. However, the
8 alkaline larger droplets in the San Joaquin Valley could lead to greater nitrate production through
9 aqueous ozonation reactions (Collett et al., 1999). The size class distinctions have substantial
10 implications for deposition of particulate pollutant species in the fog droplets due to the larger Vd
11 for impaction and occult deposition of the larger fog particles.
12 Acidic cloud water deposition has been associated with forest decline in industrialized
13 areas of the world (Anderson et al., 1999). Clouds can contain high concentrations of acids and
14 other ions. The four most prevalent ions found in cloud water samples, in decreasing order of
15 concentration, were usually sulfate (SO42"), hydrogen (H+), ammonium (NH4+), and nitrate (NO3").
16 The concentrations of these major ions tend to co-vary within within cloud events and typically
17 there was an inverse relationship between LWC of the cloud and ionic concentration of the cloud
18 water. The acidity of cloud water typically is 5 to 20 times more acid than rain water. This can
19 increase by more than 50% pollutant deposition and exposure of vegetation and soils at
20 high-elevation sites when compared with rainfall and dry deposition.
21 The widespread injury to mountain forests in document since the 1970s in West Germany
22 and other parts of Europe and more recently in the Appalachian Mountains has been attributed to
23 exposure to the cloud water reducing cold tolerance of red spruce. Forest injury also has been
24 attributed to increased leaching of cations and amino acids, increased deposition of nitrogen and
25 aluminum toxicity resulting from acidic deposition and the combined effect of acidic
26 precipitation, acid fog, oxidants, and heavy metals (Anderson et al., 1999).
27 The Mountain Acid Deposition Program (MADPro) was initiated in 1993 as part of the
28 Clean Air Status and Trends Network (CASTnet). MADPro monitoring efforts focused on the
29 design and implementation of an automated cloud water collection system in combination with
30 continuous measurement of cloud liquid water content (LWC) and meteorological parameters
31 relevant to the cloud deposition process.
April 2002 4-31 DRAFT-DO NOT QUOTE OR CITE
-------
1 Results from the MADPro automated cloud water collectors at three selected mountain
2 sites (Whiteface Mt, NY; Whitetop Mt. Va; and Clingman's Dome, TN) taken hourly from
3 nonprecipitating clouds during non-freezing seasons of the year from 1994 to 1997 were
4 promptly analyzed for pH, conductivity, and concentration of dissolved ions give an indication of
5 exposures at the three sites. Cloud liquid water content were measured at each site. The mean
6 cloud water frequencies and LWC were higher at Whiteface Mountain, NY, that in the southern
7 Appalachians. The four most prevalent ions found in cloud water samples, in order of decreasing
8 concentrations, usually were sulfate (SO42"), hydrogen (H+), ammonium (NH4+), and nitrate
9 (NO3"). The concentrations of these ions tended to co-vary within cloud events and typically
10 there was an inverse relationship between LWC of the cloud and ionic concentration of the cloud
11 water. Highest ionic concentrations were seen in mid-summer during the sampling season. Ionic
12 concentrations of samples from southern sites were significantly higher than samples from
13 Whiteface Mountain, however, further analysis indicated that this observation was due at least in
14 part to North to South differences in the LWC of clouds (Anderson et al., 1999).
15 Several factors make occult deposition particularly effective for delivery of dissolved and
16 suspended materials to vegetation. Concentrations of particulate-derived materials are often
17 many-fold higher in cloud or fog water than in precipitation or ambient air in the same area due
18 to orographic effects and from gas-liquid partitioning coefficients of specific chemical species.
19 Fog and cloud water deliver PM in a hydrated and, therefore, bioavailable form to foliar surfaces.
20 Previously dry-deposited PM may also become hydrated through delinquence or by dissolving in
21 the film of liquid water from fog deposition. The presence of fog itself maintains conditions of
22 high relative humidity and low radiation, thus reducing evaporation and contributing to the
23 persistence of these hydrated particles on leaf surfaces. Deposition of fog water is very efficient
24 (Fowler et al., 1991) with a Vd (fog 10 - 24 //m; Gallagher et al., 1988), essentially equal to the
25 aerodynamic conductance for momentum transfer (ra)4. This greatly enhances deposition by
26 sedimentation and impaction of submicron aerosol particles that exhibit very low Vd prior to fog
27 droplet formation (Fowler et al., 1989). The near equivalence of Vd and (rj"1 simplifies
28 calculation of fog water deposition and reflects the absence of vegetative physiological control
29 over surface resistance. Fog particles outside this size range may exhibit Vd below (rj"1.
30 For smaller particles, this decline reflects the increasing influence of still air and boundary layer
31 effects on impaction as particle size and momentum decline (Figure 2-1). For larger particles
April 2002 4-32 DRAFT-DO NOT QUOTE OR CITE
-------
1 momentum is sufficient to overcome these near surface limitations, but Vd may decline as
2 turbulent eddy transport to the surface becomes inefficient with increasing inertia (Gallagher
3 et al., 1988). The deposition to vegetation for PM in fog droplets is directly proportional to wind
4 speed, droplet size, concentration, and fog density (liquid water content per volume air), although
5 the latter two may be inversely related.
6 In some areas, typically along foggy coastlines or at high elevations, occult deposition
7 represents a substantial fraction of total deposition to foliar surfaces (Fowler et al., 1991,
8 Table 4-6).
9
10
TABLE 4-6. RELATIVE MAGNITUDES OF WET, DRY, AND OCCULT
DEPOSITION OF NITRATES (NO3) AND SULFATES (SO4 2) TO THREE FOREST
SITES SUBJECT TO SIMILAR GAS- AND LIQUID-PHASE POLLUTANT
CONCENTRATIONS DURING SPRING AND SUMMER3
Deposition (kg ha-1)b
Wet Dry Occult
Sitec NO3 SO42 NO3 SO42 NO3 SO42
Keilder Forest, UK
300m
Fog 11% 3 13 4 <1 17
Whitetop Mt, VA, USA
1,682m
Fog 35% 5 14 3 <1 40 120
Mt. Mitchell, NC, USA
2,006 m 5 — 3d — 18 —
aAdapted from Unsworth and Wilshaw (1989), summarizing data of Fowler et al. (1989), Mueller (1988), and
Aneja and Murthy (1994).
bAveraging periods and methods differ between sites.
°Elevation above sea level.
dOnly 0.7% of dry deposition was paniculate.
1 4.2.1.4 Magnitude of Deposition
2 Dry deposition of PM is most effective for coarse particles including primary geologic
3 material and for elements such as iron and manganese. Wet deposition is most effective for fine
4 particles of atmospheric (secondary) origin (e.g., nitrogen and sulfur, Table 4-6) and elements
April 2002 4-33 DRAFT-DO NOT QUOTE OR CITE
-------
1 such as cadmium, chromium, lead, nickel, and vanadium (Reisinger, 1990; Smith, 1990a,b,c;
2 Wiman and Lannefors, 1985). The occurrence of occult deposition is more restricted. The
3 relative magnitudes of the different deposition modes varies with ecosystem type, location,
4 elevation, and chemical burden of the atmosphere. For the Walker Branch Watershed, a
5 deciduous forest in rural eastern Tennessee, dry deposition constituted a major fraction of the
6 total annual atmospheric input of cadmium and zinc (-20%), lead («55%), and manganese
7 (-90%). Whereas wet deposition fluxes during precipitation events exceeded dry deposition
8 fluxes by one to four orders of magnitude (Lindberg and Harriss, 1981), dry deposition was
9 nearly continuous. Immersion of high-elevation forests in cloud water may occur for 10% or
10 more of the year, significantly enhancing transfer of PM and dissolved gases to the canopy.
11 Occult deposition in the Hawaiian Islands dominated total inputs of inorganic N (Heath and
12 Huebert, 1999). Much of this N was volcanically derived during the generation of volcanic fog
13 in part through reactions with seawater. In this humid climate, the dominance of occult rather
14 than wet deposition is notable.
15 High-elevation forests receive larger particulate deposition loadings than equivalent low
16 elevation sites. Higher wind speeds enhance the rate of aerosol impaction. Orographic effects
17 enhance rainfall intensity and composition and increase the duration of occult deposition.
18 Coniferous species in these areas with needle-shaped leaves also enhance impaction and
19 retention of PM delivered by all three deposition modes (Lovett, 1984).
20 In more arid regions, such as the western United States, the importance of dry deposition
21 may be larger. In the San Gabriel Mountains of southern California, for example, while annual
22 deposition of SO4"2 (partly of marine origin) was dominated by wet deposition (Fenn and Kiefer,
23 1999), deposition of NO3" was dominated by dry deposition, as was that of NH4+ at two of three
24 sites. Similarly, at a series of low elevation sites in southern California (Padgett et al., 1999), dry
25 deposition of NO3" was dominated by dry deposition. In both cases, however, the contribution of
26 gaseous HNO3 was probably substantial.
27
28 Nitrates, Sulfates and Cations
29 Much particulate sulfate and nitrate is found on particles in the 0.1- to 1.0-//m size range
30 (U.S. Environmental Protection Agency, 1982), but most of these and of base cation and heavy
31 metal inputs to forested ecosystems results from the deposition of larger particles (Lindberg and
April 2002 4-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lovett, 1985; Lindberg et al., 1982). The influence of aerodynamic diameter is particularly
2 critical for nitrogen species, because they exist as a wide range of particle sizes in the atmosphere
3 (Milford and Davidson, 1987). For example, at many sites in North America, NO3" is
4 characterized by a bimodal size distribution with modes above and below 1 //m. The
5 supermicron particles are often the result of reactions between HNO3 and coarse alkaline aerosols
6 (Wolff, 1984) as, for example, in the San Joaquin Valley of California (Lindberg et al., 1990).
7 Although the annual deposition of NH4+ is distributed similarly among the fine and coarse
8 particles, particulate NO3" is found predominantly in the coarse-particle fraction (Table 4-7).
9 Similar to the pattern for NH4+, the estimated annual deposition of SO4"2 particles occurs in both
10 the fine- and coarse-particulate fractions (Table 4-8), while base cation deposition is virtually
11 restricted to contributions from coarse particles (Table 4-9).
12 Although the annual chemical inputs to ecosystems from particle deposition is significant
13 by itself, it is important to compare it with the total chemical inputs from all sources of
14 atmospheric deposition (i.e., precipitation, particles, and gaseous dry deposition). Figure 4-6
15 shows the mean percentage contribution of NO3" andNH4+, SO4"2, and base cation-containing
16 particles to the total nitrogen, sulfur, and base cation deposition load to forest ecosystems
17 (derived from Tables 4-7 through 4-9). Although the mean contribution of particulate deposition
18 to cumulative nitrogen and sulfur deposition is typically less than 20% of annual inputs from all
19 atmospheric sources, particulate inputs of base cations average half the total base cations entering
20 forest ecosystems from the atmosphere.
21 An extensive comparison of particle to total chemical deposition is provided by the
22 Integrated Forest Study (IFS) (Johnson and Lindberg, 1992a; Lovett, 1994; Lovett and Lindberg,
23 1993; Lindberg and Lovett, 1992; Ragsdale et al., 1992). Other similar data sets are available
24 (Kelly and Meagher, 1986; Miller et al., 1993; Lindberg et al., 1986, 1990). These data in
25 (Tables 4-7 through 4-9) clearly indicate that the contribution of coarse and fine aerosols to
26 deposition to forest ecosystems is strongly dependent on the chemical species.
27 Dry deposition is an important flux of sulfur and nitrogen compounds at all of the IFS sites
28 and ranges from 9 to 59% of total (wet + dry + cloud) deposition for sulfur, 25% to 70% for
29 NO3", and, 2% to 33% for NH4+. For only NH4+ is wet deposition consistently greater than dry
30 deposition (Lovett, 1994).
31
April 2002 4-3 5 DRAFT-DO NOT QUOTE OR CITE
-------
•*tj
o
H
Q 03
Z W
~
H
W
aa
o
0
'M
(D
C^
•n os°-HOw-)Os ost^
^^ ^^ O^ O's r- H co '~H ^^ co r^j c*^ ^o
r*~) ^O ON ^O ^^ ^^ *O *O O^ I~H r*~) */^
^r,rrirf-jrf-1^Hasrs,Cf-1l>^HW-1
Os O O*s >O CN CO ^O OO 10 I-H ^ 10
^H ^H ^H ^H (N
r~--i>-^Oi>-'--|'^'c^irNcorNi>-co
rNm^o cNO'^t CN I-H
^c^OO^^oSSc^10^^
^^^SSS^S?0^10^
oq — 10
I 1 8 8 1 1*|*
S§ O tn § §0~^'c3 ^ C§ ^
K^ *U O QJ Qj | O ^^ O QJ ^^ QJ
s
"C (D (D
-------
to
tt£
•<3 H
83
e£
*0
f( ^^
HH (/}
to w
S U
og
to O
^ J
j| •<
w> ^H
•^ ^
JO
1 to S
'3 Z W
if O H
SITION (
lEPOSIT]
ECOSY!
O H ^
Cu ^y *-*
?5 H^ 2C
^ ^^
&H ^^ rV" .
H nr
^^
^^ ^^
^J ^A
P H
J H
8. MEAN ANNUA
i COMPARED TO
•* ^
~i ^^
3 u
Cfl ^^
^ S
H2
ri
s
o
o
a
Q
^H
S
^c
3
CO
"§
g
^
0
H
B
'^
a
Particle De
Coarse
&•
Q
u
.3
Location
cL
t^
a
Region/Foi
O ^O O '~H Ji- ^ fN >O "^ ^O •/") OO O Ji-* C^l
|r)-;^-p:^Hinoo^Hasn'2ri_4.fSfs2 t^os
^o r~~ 'n ^o ^~ | >o " JL ^o ro | os r —
tst-t-t-ts«nm'tI5oi,-\f^o ooos
^H^t-u-ifNfNfNmm^IJ^1^ «-i^H
S3 S2 S3
B >3 "3 "3 B
Q StDtD S 2^2^^
"tn C^ O (D TO C3 *-H "tn ^H (D >^
s°.so^^rto 0,°^,°^ | >,
iSMr-sawTSr-ar-^ja^w i>
"^ottai'^tt-S-efe'^fei Be
K2uoSS^o=gouK2uS wo
^GZHHE^SZZ^ZH o^
* u
1
II
"*
o
OS
OS
^-^
13
42
1
II
CN
OS
OS
ind Lovett (
60
flj
•§
_rt
' T
^^ r~~~
T3 OO
«i B OS
^y ?7 s
tJ OS 4D
a OS O
C 0
rt u 2
rom precipitation, gases,
Iberg (1992a), 2 = Lindb
wing season from April 1
a a P
§ j M
S||
a B -^
-------
H
8
H
f—i
<
z
s
^
o
s
2
~K
"S
c?
.>
s
•^
j^!
O
H
HH
VI
O
o-
tu
Q
^-
O
L^
g
:AN ANNUAL BASE
HH
3
^f
TABLE .
H
HH
^J
g
o
j
j
g
1
\^- rf.
O S
G w
hH H
O5 t/3
O ^
Cu ^^
uj O
Q U
^~ HH
O H
r"H ^J
•^r p^
u o
Ed ^H
t/2 HH
^ O
HH K-w
J H
WJ
,
g3
ES COMPARED TO
A^
^^
h-1
H
HH
^
PH
C^
1
Q
13
1
o
H
Is
"
,
Region/Forest T
o
r--
C3
'&
O
8
'a
"o
g
w>^t-moooo^ oo
t^^t'tritrirN^t'^H^H ^o
^H (N ^H
1 1
"S 1 ^ 1 ^ u
«, a 5 I f s i | ^
1 « 1 « ^ 1 £ I 1 1
Iglislsl si
^t
y g o o 8 Si t«
.s -s a a a | aa w
!|ll!i|l!!!
So^piPiPi
-------
Nitrate and Ammonium
Sulphate
Base Cations
Dry Particle Chemicals
Figure 4-6. Mean (±SE) percent of total nitrogen, sulfur, or base cation deposition
contributed by fine plus coarse particles. Data are means from Tables 4-7
through 4-9.
1 After emission from their sources, air pollutants are transformed and transported by
2 atmospheric processes (i.e, atmospheric meteorology) until deposited from the atmosphere to an
3 aquatic or terrestrial ecosystem. As a result, ground-level concentrations of an air pollutant
4 depend on the proximity to the sources, prevailing meteorology, and nature and extent of
5 atmospherical reactions between the source and the receptor (Holland et al., 1999). A more
6 direct relationship exists between source strength and downwind ambient concentrations for
7 primary air pollutants (e.g., SO2) than for secondary pollutants (e.g., sulfate, SO42"). Interaction
8 of the chemical and physical atmospheric processes and source locations for all of the pollutants
9 have a tendency to produce data patterns that show large spatial and temporal variability.
10 Holland et al. (1999) analyzed CASTnet monitoring data and using generative additive
11 models (GAM) estimated the form and magnitude of trends of airborne concentrations of SO2
12 SO42", and nitrogen from 1989 to 1995 at 34 rural long-term CASTnet monitoring sites in the
13 eastern United States. These models provide a highly flexible method for describing potential
April 2002
4-39
DRAFT-DO NOT QUOTE OR CITE
-------
1 nonlinear relationships between concentrations, meteorology, seasonality, and time (e.g., how
2 weekly SO2 varies as a function of temperature). For most of the 34 sites in the eastern United
3 State, estimates of change in SO2 concentrations showed a decreasing functional form in 1989-
4 1990, followed by a relatively stable period during 1991-1993), then a sharper decline beginning
5 in 1994 (Holland et al., 1999).
6 Regional trends of seasonal and annual wet deposition and precipitation-weighted
7 concentrations (PWCs) of sulfate in the United States over the period 1980-1995 were developed
8 by Shannon (1999) from monitoring date and scaled to a mean of unity. In order to reduce some
9 effects of year to year climatological variability, the unitless regional deposition and PWC trends
10 were averaged (hereafter referred to a CONCDEP). During the 16 year period examined in the
11 study, estimated aggregate emissions of SO2 in the United States and Canada fell approximately
12 12% from about 1980 to 1982, it remained roughly level for a decade and then fell approximately
13 another 15% from 1992 to 1995 — for an overall decrease of about 18%. Eastern regional trends
14 of sulfate concentrations and deposition and their average CONCDEP, also exhibited patterns of
15 initial decrease, near steady state, and final decrease with year to-to-year variability. The overall
16 relative changed in CONCDEPs are greater than the changes in SO2 emissions.
17 Concentrations and calculated deposition (concentration times amount of water) of SO4"2 at
18 the Hubbard Brook Experiment Forest (FffiEF) in the White Mountains of central New
19 Hampshire have been measured since June of 1964 (Likens et al., 2001). These measurements
20 represent the longest continuous record of precipitation chemistry in North America. The long-
21 term measurements generally concur with those of Shannon (1999) discussed above. Major
22 declines in emissions of SO2 have been observed during recent decades in the eastern United
23 States and have been correlated with significant decreases in SO4"2 concentrations in
24 precipitation (Shannon, 1999).
25 Deposition of sulfates and nitrates are very clearly linked to emissions. Reduction in
26 emissions must occur before concentrations can be reduced below current levels (Likens, et al.,
27 2001). Deposition is the key variable as sensitive ecosystems in the eastern North America have
28 not yet shown improvement in response to decreased emissions of SO2 (Driscoll et al., 1989;
29 Likens et al., 1996). Clearly, additions of other chemicals, such as nitric acid and base cations,
30 must be considered in addition to sulfur when attempting to resolve the acid rain problem
April 2002 4-40 DRAFT-DO NOT QUOTE OR CITE
-------
1 (Likens et al., 1996, 1998). The effects of sulfur and nitrogen deposition on ecosystems are
2 discussed in Section 4.2.2.2.
3 The long-term record indicates that a reduction in the deposition of basic cations (Ca+2,
4 Mg2+' K+,Na+) in bulk precipitation was associated with significant declines in sulfate deposition
5 cited above for the HBEF region (Driscoll et al., 1989). Decreases in streamwater concentrations
6 of basic cations have decreased simultaneously, suggesting that streamwater concentrations of
7 basic cations are relatively responsive to changes in atmospheric inputs. Regardless of the cause,
8 the decline in atmospheric influx of basic cations could have important effects on nutrient
9 availability as well as on the acid/base status of soil and drainage water (Driscoll et al., 1989).
10
11 Trace Elements
12 Deposition velocities for fine particles to forest surfaces have been reported in the range of
13 1 to 15 cm s"1 (Smith, 1990a). For example, total, annual heavy metal deposition amounts are
14 highly variable depending on specific forest location and upwind source strength (Table 4-10).
15 Lindberg et al. (1982) quantified the dry deposition of heavy metals to inert surfaces and to
16 leaves of an upland oak forest. As noted for other chemical species, Vd was highly dependent on
17 particle size and chemical species (Table 4-11) with the larger particles depositing more
18 efficiently.
19
20
TABLE 4-10. MEAN (±SE) PARTICLE SIZE, DEPOSITION RATES, AND DERIVED
DEPOSITION VELOCITIES (Vd) FOR HEAVY METAL DEPOSITION TO THE
UPPER CANOPY (INERT PLATES OR LEAVES) OF AN UPLAND OAK FOREST
Metal
Manganese
Cadmium
Zinc
Lead
Particle Size (//m)
3.4 ±0.7
1.5 ±0.7
0.9 ±0.2
0.5
Deposition Rate
(pg cm'2 h'1)
91 ±23
0.3 ±0.1
6± 1
23 ±8
vd
(cm s"1)
6.4 ±3. 6
0.37±0.18
0.38 ±0.1
0.06 ±0.01
Source: Lindberg etal. (1982).
April 2002 4-41 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-11. TOTAL HEAVY METAL DEPOSITION TO TEMPERATE
LATITUDE FORESTS
Forest Deposition
Heavy Metal kg ha"1 year"1 (Range)
Cadmium 0.002-0.02
Copper 0.016 - 0.24
Lead 0.099 - 1000
Nickel 0.014-0.15
Zinc 0.012-0.178
Source: Smith (1990c).
1 Preferential association of heavy metals with fine particles allows these particles to escape
2 emission controls. Metal removal efficiencies for baghouse filters are typically 95 to 99% for all
3 but mercury, but fine particle capture is much less efficient. Wet scrubber efficiency varies with
4 design and pressure drop, typically 50 to 90% (McGowan et al., 1993). Fine particles also have
5 the longest atmospheric residence times and, therefore, can be carried long distances. Depending
6 on climate conditions and topography, fine particles may remain airborne for days to months and
7 may be transported 1,000 to 10,000 km or more from their source. This long-distance transport
8 and subsequent deposition qualify heavy metals as regional- and global-scale air pollutants.
9 Ecosystems immediately downwind of major emissions sources (such as power generating,
10 industrial, or urban complexes) may receive locally heavy inputs. Mass balance budgets (inputs
11 and outputs) of seven heavy metals (cadmium, copper, iron, lead, manganese, nickel, and zinc)
12 have been determined at the Hubbard Brook Experimental Forest (White Mountain National
13 Forest) in New Hampshire. This forest is about 120 km northwest of Boston and relatively
14 distant from major sources of heavy metal emissions. However, continental air masses that have
15 passed over centers of industrial and urban activity also frequently follow storm tracks over
16 northern New England. The resulting annual input for the seven heavy metals at Hubbard Brook
17 for 1975 to 1991 is presented in Table 4-12. Inputs of most of the heavy metal species remained
18 relatively constant over the 16-year period. The 44-fold decrease in lead deposition is correlated
19 with removal of lead from motor vehicle fuels.
April 2002 4-42 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-12. ANNUAL3 BULK DEPOSITION" INPUT OF SEVEN HEAVY METALS
TO THE HUBBARD BROOK EXPERIMENTAL FOREST (43° 56' N LATITUDE,
71° 45' W LONGITUDE), WHITE MOUNTAIN NATIONAL FOREST, NH, FOR THE
PERIOD 1975 TO 1991C (grams per hectare)
1975
1976
1977
1978
1979
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
Cadmium
2.5
3.0
40.0
11.0
16.0
8.0
8.0
7.5
6.3
4.5
2.3
2.2
1.6
2.5
2.1
Copper
18.2
11.6
10.0
26.0
16.0
14.0
17.0
18.6
9.0
6.8
4.5
4.7
3.3
10.5
15.0
Iron
832.0
1,214.0
372.0
234.0
207.0
178.0
206.0
217.0
174.0
128.0
16.0
145.0
160.0
124.0
134.0
Lead
352.0
359.0
195.0
141.0
155.0
70.0
57.0
56.6
41.0
25.9
17.2
12.5
11.9
11.8
8.3
Manganese
100.0
199.0
39.0
74.0
172.0
49.0
52.0
85.0
53.8
58.4
55.8
65.6
74.2
42.4
72.0
Nickel
29.4
18.0
8.0
13.0
12.0
13.0
7.5
7.7
7.0
9.4
10.7
8.2
7.4
8.0
9.6
Zinc
175.0
182.0
116.0
95.0
278.0
54.0
76.0
73.8
54.8
54.7
45.2
47.0
57.2
47.1
55.5
aMean of monthly totals.
bTotal input, including both wet and dry deposition.
cData not available for 1980 and 1981.
1 Trace element investigations conducted in roadside, industrial, and urban environments
2 have demonstrated that impressive burdens of particulate heavy metals accumulate on vegetative
3 surfaces. Lead deposition to roadside vegetation (prior to its removal from fuel) was 5 to 20,
4 50 to 200, and 100 to 200 times lead deposition to agricultural crops, grasses, and trees,
5 respectively, in non-roadside environments. In an urban setting, it has been estimated that the
6 leaves and twigs of a 30-cm (12-in.) diameter sugar maple remove 60, 140, 5800, and 820 mg of
7 cadmium, chromium, lead, and nickel, respectively, during the course of a single growing season
8 (Smith, 1973).
April 2002 4-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 Effective deposition of PM is required before biological effects on plants or ecosystems can
2 occur. It is clear that substantially improved techniques for monitoring and predicting deposition
3 will be required to characterize these effects with certainty.
4
5 Semivolatile Organics
6 Organic compounds partition between gas and particle phases, and particulate deposition
7 depends largely on the particle sizes available for adsorption (Pankow, 1987; Smith and Jones,
8 2000). Dry deposition of organic materials (eg. dioxins, dibenzofurans, polycyclic aromatics) is
9 often dominated by the coarse fraction, even though mass loading in this size fraction may be
10 small (Lin et al., 1993) relative to the fine PM fraction. For example, measurements in Bavaria
11 in both summer and winter revealed that >80% of organics were in the fine (<1.35 //m) fraction
12 (Kaupp and McLachlan, 1999). Nevertheless, in most cases, calculated values of dry deposition
13 were dominated by the material adsorbed to coarse particles. Wet deposition, in contrast, was
14 dominated by the much larger amount of material associated with fine particles. In this
15 environment (where monthly precipitation is about 50 mm in winter and summer), wet
16 deposition dominated, with dry deposition accounting for only 14 to 25% of total deposition
17 (Kaupp and McLachlan, 1999). Lower relative contents of more volatile species in summer than
18 winter (Kaupp and McLachlan, 1999) indicate the critical importance of gas-particle phase
19 interconversions in determining deposition.
20
21 4.2.2 Effects on Vegetation and Ecosystems
22 Exposure to a given mass concentration of airborne PM may lead to widely differing
23 phytotoxic responses, depending on the particular mix of deposited particles. The most common
24 and useful subdivision of PM, derived from the typical bimodal distribution of atmospheric
25 particles, is into fine and coarse particles (Wilson and Suh, 1997). The smallest particle at or
26 near 1.0 to 2.5 //m generally is taken as the division between fine and coarse, although this is not
27 an absolute and is subject to some shift (e.g., with changing ambient humidity). However, the
28 typical the rule of thumb, as used in the 1996 PM AQCD (U.S. Environmental Protection
29 Agency, 1996a), is that fine PM nominally falls in the range of 0 to 2.5 //m and coarse-mode PM
30 in the range of 2.5 to 10.0 //m.
April 2002 4-44 DRAFT-DO NOT QUOTE OR CITE
-------
1 Effects of particulate deposition on individual plants or ecosystems are difficult to
2 characterize because of the complex interactions among biological, physicochemical, and
3 climatic factors. Most direct effects other than regional effects associated with global changes
4 occur in the severely polluted areas surrounding industrial point sources, such as limestone
5 quarries, cement kilns, and metal smelting facilities. Fine particles have greater distribution.
6 Experimental applications of PM constituents to foliage typically elicit little response at the more
7 common ambient concentrations. The diverse chemistry and size characteristics of ambient PM
8 and the lack of clear distinction between effects attributed to phytotoxic particles and to other air
9 pollutants further confound understanding of the direct effects on foliar surfaces. The majority of
10 documented toxic effect of particles on vegetation reflect their chemical content (acid/base, trace
11 metal, nutrient), surface properties, or salinity. Studies of direct effects of particles on vegetation
12 have not yet advanced to the stage of reproducible exposure experiments. Difficulties in
13 experimental application of ambient particles to vegetation have been discussed by Olszyk et al.
14 (1989). It is now clear that many phytotoxic gases are deposited more readily, assimilated more
15 rapidly, and lead to greater direct injury to vegetation than do most common particulate materials
16 (Guderian, 1986). The dose-specific responses (dose-response curves) obtained in early
17 experiments following the exposure of plants to phytotoxic gases generally have not been
18 observed following the application of particles.
19 Unlike gaseous dry deposition, neither the solubility of the particles nor the physiological
20 activity of the surface is likely to be of first order of importance in determining deposition
21 velocity (Vd). Factors that contribute to surface wetness and stickiness may be critical
22 determinants of sticking efficiency. Available tabulation of deposition velocities are highly
23 variable and suspect. High-elevation forests receive larger particle deposition loadings than
24 equivalent lower elevations sites because of higher wind speeds and enhanced rates of aerosol
25 impaction; orographic effects on rainfall intensity and composition; increased duration of occult
26 deposition; and, in many areas, the dominance of coniferous species with needle-shaped leaves
27 (Lovett, 1984). Recent evidence indicates that all three modes of deposition (wet, occult, and
28 dry) must be considered in determining inputs to ecosystems or watersheds, because each may
29 dominate over specific intervals of space.
30 Coarse-mode particles are primary in nature, having been produced and emitted from a
31 point or area source as a fully formed particle. They generally range in size from ca. 2.5 to
April 2002 4-45 DRAFT-DO NOT QUOTE OR CITE
-------
1 100 (j,m. This material is created by abrasion and may be subsequently suspended by wind or
2 mechanical means. Suspended geologic material contains the chemical and, potentially, the
3 biological signature of the soil from which it derives (e.g., it may be dominated by iron, silica,
4 aluminum, and/or calcium). Additional anthropogenically derived coarse-mode PM derives from
5 fly ash, automobile tires and brake linings, and industrial effluent associated with crushing and
6 grinding operations. Coarse-mode particles also include biogenically derived organic materials
7 (e.g., fragments of plants and insects, pollen, fungal spores, bacteria, and viruses in marine
8 aerosols).
9 In general, fine-mode PM is secondary in nature, having condensed from the vapor phase or
10 been formed by chemical reaction from gaseous precursors in the atmosphere. These particles
11 exist in a nucleation mode (having a mass median aerodynamic diameter or MMAD of about
12 0.06 //m) and may grow by coagulation of existing particles or by condensation of additional
13 gases onto existing particles into an accumulation mode (about 0.5 //m). Sulfur and nitrogen
14 oxides (SOX and NOX), as well as volatile organic gases, are common precursors for fine PM and
15 are often neutralized with ammonium cations as particulate salts. Condensation of volatilized
16 metals and products of incomplete combustion also are common precursors. Reactions of many
17 of these materials with an oxidizing atmosphere lead to high secondary PM concentrations during
18 the summer months in many parts of the United States.
19 Atmospheric PM may affect vegetation directly following deposition on foliar surfaces or
20 indirectly by changing the soil chemistry or through changes in the amount of radiation reaching
21 the Earth's surface through PM-induced climate change processes. Indirect effects, however, are
22 usually the most significant because they can alter nutrient cycling and inhibit plant nutrient
23 uptake. The possible direct responses to PM deposition are considered in this section, and the
24 indirect responses are discussed in the later sections on ecosystems.
25
26 4.2.2.1 Direct Effects of Particulate Matter on Individual Plant Species
27 Particles transferred from the atmosphere to foliar surfaces may reside on the leaf, twig, or
28 bark surface for extended periods; be taken up through the leaf surface; or be removed from the
29 plant via resuspension to the atmosphere, washing by rainfall, or litter-fall with subsequent
30 transfer to the soil. Any PM deposited on above-ground plant parts may exert physical or
31 chemical effects. The effects of "inert" PM are mainly physical; whereas those of toxic particles
April 2002 4-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 are both chemical and physical. The effects of dust deposited on plant surfaces or soil are more
2 likely to be associated with their chemistry than simply with the mass of deposited particles and
3 may be more important than any physical effects (Farmer, 1993). Nevertheless, vegetative
4 surfaces represent filtration and reaction/exchange sites (long, 1991; Youngs et al., 1993).
5
6 Effects of Coarse Particles
7 Coarse particles, ranging in size from 2.5 to 100 //m, are chemically diverse, dominated by
8 local sources, and typically deposited near their source because of their sedimentation velocities.
9 Airborne coarse particles are derived from road, cement kiln, and foundry dust; fly ash; tire
10 particles and brake linings; soot and cooking oil droplets; biogenic materials (e.g, plant pollen,
11 fungal spores, bacteria and viruses); abraded plant fragments; sea salt; and hydrated deliquescent
12 particles of otherwise fine aerosol. In many rural areas and some urban areas, the majority of
13 mass in the coarse particle mode is in the elements silicon, aluminum, calcium, and iron,
14 suggesting a crustal origin as fugitive dust from disturbed land, roadways, agriculture tillage, or
15 construction activities. Rapid sedimentation of coarse particles tends to restrict their direct
16 effects on vegetation largely to roadsides and forest edges.
17
18 Physical Effects—Radiation. Dust can cause physical and chemical effects. Deposition of
19 inert PM on above-ground plant organs may result in an increase in radiation received, a rise in
20 leaf temperature and the blockage of stomata. Increased leaf temperature, heat stress, reduced net
21 photosynthesis, and leaf chlorosis, necrosis, and abscission were reported by Guderian (1986).
22 Road dust decreased the leaf temperature on Rhododendron catawbiense by ca. 4 °C (Eller,
23 1977); whereas foundry dust caused an 8.7 °C increase in leaf temperature of black poplar
24 (Populus nigra) under the conditions of the experiment (Guderian, 1986). Deciduous (broad)
25 leaves exhibited larger temperature increases because of particle loading than did conifer (needle)
26 leaves, a function of poorer coupling to the atmosphere. Inert road dust caused a three- to four-
27 fold increase in the absorption coefficient of leaves of English ivy (Eller, 1977; Guderian, 1986)
28 for near infrared radiation (NIR; 750 to 1350 nm). Little change in absorption occurred for
29 photosynthetically active radiation (PAR; 400 to 700 nm). The increase in NIR absorption was
30 equally at the expense of reflectance and transmission in these wavelengths. The net energy
31 budget increased by ca. 30% in the dust-affected leaves. Deposition of coarse particles increased
April 2002 4-47 DRAFT-DO NOT QUOTE OR CITE
-------
1 leaf temperature and contributed to heat stress; reduced net photosynthesis; and caused leaf
2 chlorosis, necrosis, and abscission (Dassler et al., 1972; Parish, 1910; Guderian, 1986; Spinka,
3 1971).
4 Starch storage in dust-affected leaves increased with dust loading under high (possibly
5 excessive) radiation, but decreased following dust deposition when radiation was limiting. These
6 modifications of the radiation environment had a large impact on single-leaf utilization of light.
7 The boundary layer properties, determined by leaf morphology and environmental conditions,
8 strongly influenced the direct effects of particle deposition on radiation heating (Eller, 1977;
9 Guderian, 1986) and on gas exchange as well. Brandt and Rhoades (1973) attributed the
10 reduction in growth of trees to crust formation from limestone dust on the leaves. Crust
11 formation reduced photosynthesis and the formation of carbohydrates needed for normal growth,
12 induced premature leaf-fall, damaged leaf tissues, inhibited growth of new tissue, and reduced
13 starch storage. Dust may decrease photosynthesis, respiration, and transpiration; and it may
14 allow penetration of phytotoxic gaseous pollutants, thereby causing visible injury symptoms and
15 decreased productivity. Permeability of leaves to ammonia increased with increasing dust
16 concentrations and decreasing particle size (Farmer, 1993).
17 Dust also has been reported to physically block stomata (Krajickova and Mejstfik, 1984).
18 Stomatal clogging by particulate matter from automobiles, stone quarries, and cement plants was
19 also studied by Abdullah and Iqbal (1991). The percentage of clogging was low in young leaves
20 when compared with old and mature leaves and the amount of clogging varied with species and
21 locality. The maximum clogging of stomata observed was about 25%. The authors cited no
22 evidence that stomatal clogging inhibited plant functioning. The heaviest deposit of dust is
23 usually on the upper surface of broad-leaved plants; whereas the majority of the stomata are on
24 the lower surface where stomatal clogging would be less likely.
25
26 Chemical Effects. The chemical composition of PM is usually the key phytotoxic factor
27 leading to plant injury. Cement-kiln dust on hydration liberates calcium hydroxide that can
28 penetrate the epidermis and enter the mesophyll; and, in some cases, this has caused the leaf
29 surface alkalinity to reach a pH of 12. Lipid hydrolysis, coagulation of the protein compounds,
30 and ultimately plasmolysis of the leaf tissue result in reduction in the growth and quality of plants
31 (Guderian, 1986). In experimental studies, application of cement kiln dust of known
April 2002 4-48 DRAFT-DO NOT QUOTE OR CITE
-------
1 composition for 2 to 3 days yielded dose-response curves between net photosynthetic inhibition
2 or foliar injury and dust application rate (Darley, 1966). Lerman and Darley (1975) determined
3 that leaves must be misted regularly to produce large effects. Alkalinity was probably the
4 essential phytotoxic property of the applied dusts.
5
6 Salinity. Particulate matter enters the atmosphere from oceans following the mixing of air
7 into the water column and the subsequent bursting of bubbles at the surface. The effervescence
8 of bubbles on the surface of the ocean forcefully ejects droplets of sea water into the air. These
9 droplets, concentrated by evaporation, are carried inland by wind and deposited on the seaward
10 side of coastal plants (Boyce, 1954). This occurs largely in the surf line (i.e., near land and
11 potentially sensitive terrestrial receptors). This process can be a significant source of sulfate,
12 sodium, chloride, and trace elements (as well as living material) in the atmospheric aerosol
13 impacting coastal vegetation. Sea-spray particles (Taback et al., 1979) are approximately 24%
14 greater in size than 10 //m, and 54% are between 3 and 10 //m. Thus, approximately only 20%
15 are fine (0 to 2.5 //m) particles; and deposition by sedimentation and impaction is concentrated
16 near the coast. Airborne concentrations of this marine PM decrease quickly with distance inland
17 from the surfline both by deposition and dilution within atmospheric mixed layer (McKay et al.,
18 1994; Nelis et al., 1994). Near-shore sediments with associated pollutants present in coastal
19 runoff may be suspended in the surf and reentrained into the air. This can be a substantial source
20 of microorganisms and of radionuclides to coastal vegetation (Nelis et al, 1994; McKay et al.,
21 1994).
22 Sea-salt particles can serve as nuclei for the absorption and subsequent reaction of other
23 gaseous and paniculate air pollutants. Both nitrate and sulfate from the atmosphere have been
24 found to associate with coarse and fine sea-salt particles (Wu and Okada, 1994). Direct effects
25 on vegetation reflect these inputs, as well as classical salt injury caused by the sodium and
26 chloride that constitute the bulk of these particles. Foliar accumulation of airborne salt particles
27 may lead to foliar injury, thusly affecting the species composition in coastal environments
28 (Smith, 1984).
29 The effects of winds and sea spray on coastal vegetation has been reported in the literature
30 since the early 1800s (Boyce, 1954). However, there has been a difference of opinion as to
31 whether the injury to coastal vegetation resulted from windblown aerial salts or from mechanical
April 2002 4-49 DRAFT-DO NOT QUOTE OR CITE
-------
1 injury (i.e., sand blasting) due to wind alone. Though the significance of sea water dashed on
2 fore dunes and rocky coasts had been recognized by several authors, Wells and Shunk (1937,
3 1938) and Wells (1939) were the first to recognize the importance of salt spray in coastal
4 ecology. Wells and Shunk (1937) reported that salt spray carried over dunes was the most
5 important factor influencing growth form, zonation, and succession in coastal dunes. Salt spay
6 injury was recorded 1.25 miles inland on the North Carolina coast. On the basis of observations
7 in the Cape Fear area, they determined that the shape of coastal "wind form" shrubs were the
8 result of sea spray carried by high winds. They found injury on shrubs only near the coast while
9 those at a greater distances inland showed no injury whatsoever after a strong southeast wind that
10 persisted for a period of nineteen hours during cloudy weather and abundant soil moisture.
11 To determine the cause of injury, injured and uninjured shoots were titrated for chlorides.
12 A marked difference was observed between the injured and uninjured shoots (Wells and Shunk,
13 1937, 1938). Experimental spraying of shoots of woody plants with seawater resulted in a
14 pattern of injury similar to the injury observed on seaside shrubs. The absence of the more inland
15 species, such as persimmon (Diosporos virginiana L.), turkey oak (Quercus laevis Walt.),
16 longleaf pine (Pinuspalustris Mill., P. australis Michx.), and wire grass (Aristida stricta
17 Michx.), was explained on the basis of intolerance of these species to salt spray. The dominance
18 of live oak (Quercus virginiana Mill.), as a practically pure stand on Smith Island (also known as
19 Bald Head Island), NC and along the eastern and southern NC coast, was determined by Wells
20 (1939) to be due to the tree's tolerance to salt spray. He termed the long term stabilization of the
21 live oak stand as a new type of climax, the "salt spray climax." The later work of Costing and
22 Billings (1942) near Beaufort, NC corroborated the findings of Wells and Shunk, 1937, 1938).
23 The report by Boyce (1954) is probably the most extensive on salt-spray communities.
24 Dune sands in many coastal areas have been shown to have extremely low concentrations of
25 dissolved salts. Studies have indicated that the salt content of the coastal dunes of Virginia,
26 Massachusetts, and California did not exceed the maximum occurring in ordinary cultivated
27 soils. Costing and Billings (1942) found no correlation between soil salinity and plant
28 distribution on the North Carolina coast. Surface crusts of sand dunes have been shown to have
29 high concentrations of chlorides which could be attributed to sea spray, while concentrations of
30 chlorides in underlying layers was low. The surface layer, however, varied with exposure of the
31 dunes to oceanic winds (Boyce, 1954).
April 2002 4-50 DRAFT-DO NOT QUOTE OR CITE
-------
1 Boyce (1954), Wells (1939), and Wells and Shunk (1938) concluded on the basis of their
2 studies that necrosis and death of plant tissues results from the high deposition of salt spray and
3 high accumulation of the chloride ion in the plant tissues. Very little salt is taken up by plant
4 roots; most enters through the aerial organs. Leaves of plants exposed to salt spray show a
5 distinct pattern of injury (Wells and Shunk, 1938). Necrotic areas first appear at the leaf tips and
6 upper margins and then progress slowly in an inverted "V" toward the petiole. This leaf injury
7 pattern was verified experimentally. Mechanical injury resulting from leaves and twigs beating
8 against each another in the wind causes the formation of small lesions through which salt can
9 enter. After entry into the plant, the chloride ion is rapidly translocated to the apices of the leaves
10 and twigs where it accumulates to injurious concentrations and results in the death of only a
11 portion of the plant. The differential deposition and translocation of the chloride ion results in
12 the death of the seaward leaves and twigs. The result is the continued growth of the uninjured
13 branches in an inland direction. As a result, the canopy angle varies with the intensity of the
14 spray (Boyce, 1954).
15 Little or no mineral ions are available in the silicate sands of the of coastal dunes. As a
16 consequence, plants obtain the mineral ions needed for growth from the salt spray. Seawater
17 contains all of the mineral ions required for growth, except nitrogen and phosphorus. The
18 amount of nitrogen and phosphorus in seawater varies over a wide range (Boyce, 1954).
19 Experiments indicated that available nitrogen in sea spray was a conditioning factor. Low
20 nitrogen availability increased the tolerance of dune species to salt spray. Increasing the
21 availability of nitrogen resulted in a different pattern of plant zonation and distribution.
22 Dicotyledonous species were restricted to areas of lower spray intensity. The severity of chloride
23 injury was associated more with the amount of available potassium than with the concentration
24 of chlorides within the limits of 280-360 mg Cl/liter (Boyce, 1954).
25 Other sources of phytotoxic saline PM include aerosols from cooling towers and roadway
26 deicing salt. Cooling towers used to dissipate waste heat from steam-electric power generating
27 facilities may emit salt if brackish water or saltwater is used as a coolant (McCune et al., 1977;
28 Talbot, 1979). Foliar injury is related to salt droplets deposited by sedimentation or impaction
29 from cooling tower drift. The distance of the salt drift determines the amount of deposition and
30 location of injury. Environmental conditions most conducive to injury were absence of
31 precipitation, which can wash salt off leaves, and high relative humidity (RH; Talbot, 1979).
April 2002 4-51 DRAFT-DO NOT QUOTE OR CITE
-------
1 Increased injury is associated with wind speed and salt concentrations. Typical toxicity
2 symptoms from acute exposures include marginal foliar necrosis and lesions; shoot-tip dieback;
3 leaf curl; and interveinal necrosis (McCune et al., 1977). Based on experimental data, Grattan
4 et al. (1981) observed that, to cause injury, salt deposited on leaf surfaces must dissolve and be
5 absorbed into leaf tissue. Their work also indicated the importance of RH in foliar uptake. If RH
6 remained below 70%, even heavy deposition of salt did not induce injury in peppers, soybeans,
7 and tomatoes.
8 Injury to vegetation from the application of deicing salt was related to salt spray blown or
9 drifting from the highways by Hofstra and Hall (1971) and Viskari and Karenlampi (2000). The
10 most severe injury was observed nearest to the highways. The results presented in these studies
11 agrees with that of Wyttenbach et al. (1989), who observed that conifers planted near roadway
12 margins in the eastern United States often exhibit foliar injury due to toxic levels of saline
13 aerosols deposited from deicing solutions. Piatt and Kranse (1974) demonstrated that road and
14 site factors influence the spread of deicing salt into forested areas. The slope away from the road
15 influenced the distance from the road where injury was observed. The percent slope was
16 correl ated with the di stance.
17
18 Effects of Fine Particles
19 Fine PM is generally secondary in nature, having condensed from the vapor phase or been
20 formed by chemical reaction from gaseous precursors in the atmosphere, and is generally smaller
21 than 1 to 2.5 //m. Nitrogen and sulfur oxides, volatile organic gases, condensates of volatilized
22 metals, and products of incomplete combustion are common precursors for fine PM. Reactions
23 of many of these materials with an oxidizing atmosphere contribute to high secondary PM
24 concentrations during summer months in many U.S. areas. The conclusion reached in the 1982
25 PM AQCD (U.S. Environmental Protection Agency, 1982), that sufficient data were not
26 available for adequate quantification of dose-response functions for direct effects of fine aerosols
27 on vegetation, continues to be true today. Only a few studies on the direct effects of acid
28 aerosols have been completed (U. S. Environmental Protection Agency, 1982). The major
29 effects are indirect and occur through the soil (Section 4.3).
30
April 2002 4-52 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nitrogen. Nitrate is observed in both fine and coarse particles. Nitrates from atmospheric
2 deposition represent a substantial fraction of total nitrogen inputs to southeastern forests (Lovett
3 and Lindberg, 1986). However, much of this is contributed by gaseous nitric acid vapor, and a
4 considerable amount of the particulate nitrate is taken up indirectly through the soil. Garner et al.
5 (1989) estimated deposition of nitrogen to forested landscapes in eastern North America at 10 to
6 55 kg/ha/year for nitrate and 2 to 10 kg/ha/year for ammonium. About half of these values were
7 ascribed to dry deposition.
8 Atmospheric additions of particulate nitrogen in excess of vegetation needs are lost from
9 the system, mostly as leachate from the soil as nitrate. Managed agricultural ecosystems may be
10 able to utilize deposited particulate nitrogen more efficiently than native ecosystems, although
11 many cultivated systems also lose considerable nitrogen as nitrate in runoff, deep drainage, or
12 soil water. It has proven difficult to quantify direct foliar fertilization by uptake of nitrogen from
13 ambient particles.
14 There is no doubt that foliar uptake of nitrate can occur, as clearly shown by the efficacy of
15 foliar fertilization in horticultural systems. Potassium nitrate was taken up by leaves of
16 deciduous fruit trees (Weinbaum and Neumann, 1977) and resulted in increased foliar nitrogen
17 concentrations. Not all forms of nitrogen are absorbed equally, nor are all equally benign.
18 Following foliar application of 2600 ppm of nitrogen as Ca(NO3)2, (NH4)2SO4, or (NH2)2CO to
19 apple canopies (Rodney, 1952; Norton and Childers, 1954), leaf nitrogen levels were observed to
20 increase to similar levels; but calcium nitrate and ammonium sulfate caused visible foliar injury,
21 whereas urea did not. Urea is generally the recommended horticultural foliar fertilizer.
22 The mechanism of uptake of foliarly deposited nitrate is not well established. Nitrate
23 reductase is generally a root-localized enzyme. It is generally not present in leaves, but is
24 inducible there. Induction typically occurs when the soil is heavily enriched in NO3". As the root
25 complement of nitrate reductase becomes overloaded, unreduced nitrate reaches the leaves
26 through the transpiration stream. Nitrate metabolism has been demonstrated in leaf tissue
27 (Weinbaum and Neumann, 1977) following foliar fertilization. Residual nitrate reductase
28 activity in leaves may be adequate to assimilate typical rates of particulate nitrate deposition.
29 Uptake of nitrate may be facilitated by codeposited sulfur (Karmoker et al., 1991; Turner and
30 Lambert, 1980).
April 2002 4-53 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nitrate reductase is feedback-inhibited by its reaction product NH4+. The common
2 atmospheric aerosol, NH4NO3, therefore may be metabolized in two distinct biochemical steps:
3 first the ammonium (probably leaving nitric acid) and then the nitrate. Losses of nitric acid by
4 volatization during this process, if they occur, have not been quantified.
5 Direct foliar effects of particulate nitrogen have not been documented. Application of a
6 variety of fine nitrogenous aerosol particles (0.25 //m) ranging from 109 to 244 //g/m3 nitrogen
7 with or without 637 //g/m3 sulfur caused no consistent short-term (2- to 5-h) effect on gas
8 exchange in oak, maize, or soybean leaves (Martin et al., 1992).
9 Although no evidence exists for the direct transfer of nutrient parti culate aerosols into
10 foliage, a few studies give insights into the potential for ammonium and nitrate transfer into
11 leaves. Fluxes of both NO3" and NH4+, measured in wet deposition and in throughfall plus
12 stemflow in forests, commonly indicate higher fluxes of nitrogen above the canopy (Parker,
13 1983; Lindberg et al., 1987; Sievering et al., 1996), indicating net foliar uptake. Lovett and
14 Lindberg (1993) reported a linear relationship between inorganic nitrogen fluxes in deposition
15 and throughfall, suggesting that uptake may be considered passive to some extent.
16 Garten and Hanson (1990) studied the movement of 15N-labeled nitrate and ammonium
17 across the cuticles of red maple (Acer rubruni) and white oak (Quercus alba) leaves when
18 applied as an artificial rain mixture. Brumme et al. (1992), Bowden et al. (1989), and Vose and
19 Swank (1990) have published similar data for conifers. These studies show the potential for
20 nitrate and ammonium to move into leaves, where it may contribute to normal physiological
21 processes (e.g., amino acid production; Wellburn, 1990). Garten (1988) showed that internally
22 translocated 35S was not leached readily from tree leaves of yellow poplar (Liriodendron
23 tulipifera) and red maple (Acer rubrum\ suggesting that SO4"2 would not be as mobile as the
24 nitrogen-containing ions discussed by Garten and Hanson (1990). Further, when the foliar
25 extraction method is used, it is not possible to distinguish sources of chemicals deposited as
26 gases or particles (e.g., nitric acid [HNO3], nitrogen dioxides [NO2], nitrate [NO3"]), or sources of
27 ammonium (deposited as ammonia [NH3] or ammonium ion [NH4+]) (Garten and Hanson, 1990).
28 Particle deposition contributes only a portion of the total atmospheric nitrogen deposition
29 reaching vegetation; but, when combined with gaseous and precipitation-derived sources, total
30 nitrogen deposition to ecosystems has been identified as a possible causal factor leading to
31 changes in natural ecosystems (See Section 4.2.3).
April 2002 4-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sulfur. Anthropogenic sulfur emissions are >90% SO2. Most of the remaining emission of
2 sulfur is directly as sulfate (U.S. Environmental Protection Agency, 1996a). Sulfur dioxide is
3 hydrophilic and is rapidly hydrated and oxidized to sulfite and bisulfite and then to sulfate, which
4 is approximately 30-fold less phytotoxic. The ratio of SO4"2/SO2 increases with aging of the air
5 mass and, therefore, with distance from the source. Sulfate is sufficiently hygroscopic in humid
6 air that it may exist significantly in the coarse particulate fraction. Because dilution of both SO2
7 and paniculate SO4"2 occurs with distance from the source, it is unusual for damaging levels of
8 parti culate sulfate to be deposited. Gas to particle conversion in this case is of benefit to
9 vegetation.
10 Sulfur is an essential plant nutrient. Low dosages of sulfur serve as a fertilizer, particularly
11 for plants growing in sulfur-deficient soil (Hogan et al., 1998). However, current levels of
12 sulfate deposition reportedly exceed the capacity of most vegetative canopies to immobilize the
13 sulfur (Johnson, 1984). Nitrogen uptake in forests may be regulated loosely by sulfur
14 availability, but sulfate additions in excess of needs do not typically lead to injury (Turner and
15 Lambert, 1980).
16 There are few field demonstrations of foliar sulfate uptake (Krupa and Legge, 1986, 1998).
17 Sulfate in throughfall is often enriched above levels in precipitation. The relative importance of
18 foliar leachate and prior dry-deposited sulfate particles remains difficult to quantify (Cape et al.,
19 1992). Leaching rates are not constant and may respond to levels of other pollutants, including
20 acids. Uptake and foliar retention of gaseous and particulate sulfur are confounded by variable
21 rates of translocation and accessibility of deposited materials to removal and quantification by
22 leaf washing. Following soil enrichment with 35SO4"2 in a Scots pine forest, the apparent
23 contribution of leachate to throughfall was only a few percent following an initial burst of over
24 90% because of extreme disequilibrium in labeling of tissue sulfate pools (Cape et al., 1992).
25 Olszyk et al. (1989) provide information on the effects of multiple pollutant exposures
26 including particles (NO3~, 142 //g/m3; NH4+, 101 //g/m3; SO4'2, 107 //g/m3). They found that only
27 gaseous pollutants produced direct (harmful) effects on vegetation for the concentrations
28 documented, but the authors hypothesized that long-term accumulation of the nitrogen and sulfur
29 compounds contributed from particle deposition might have effects on plant nutrition over long
30 periods of time. Martin et al. (1992) exposed oak (Quercus macrocarpa), soybean (Glycine
31 max}, and maize (Zea mays) plants to acute exposures (2 to 5 h) of aerosols (0.25 //m) containing
April 2002 4-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 only nitrate (109 //g/m3), ammonium and nitrate (244 and 199 //g/m3), or ammonium and sulfate
2 (179 and 637 //g/m3). They found that these exposures, which exceeded the range of naturally
3 occurring aerosol concentrations, had little effect on foliar photosynthesis and conductance.
4 Martin et al. (1992) concluded that future investigations should focus on the effects of particles
5 on physiological characteristics of plants following chronic exposures.
6
7 Acidic Deposition. The effects of acidic deposition have been accorded wide attention in
8 the media and elsewhere (Altshuller and Linthurst, 1984; Hogan et al., 1998). Probably the most
9 extensive assessment of acidic deposition processes and effects is the NAPAP Biennial Report to
10 Congress: An Integrated Assessment (National Science and Technology Council, 1998).
11 Concern regarding the effects of acidic deposition on crops and forest trees has resulted in
12 extensive monitoring and research. Exposures to acidic rain or clouds can be divided into
13 "acute" exposures to higher ionic concentrations (several //mol/L) and "chronic" long-term
14 repeated exposures to lower concentrations (Cape, 1993). Pollutant concentrations in rainfall
15 have been shown to have little capacity for producing direct effects on vegetation (Altshuller and
16 Linthurst, 1984; Hogan et al., 1998); however, fog and clouds, which may contain solute
17 concentrations up to 10 times those found in rain, have the potential to cause direct effects. More
18 than 80% of the ionic composition of most cloud water is made up of four major pollutant ions:
19 H+, NH4+, NO3", and SO4"2. Ratios of hydrogen to ammonium and sulfate to nitrate vary from site
20 to site with all four ions usually present in approximately equal concentrations. Available data
21 from plant effect studies suggest that hydrogen and sulfate ions are more likely to cause injury
22 than ions containing nitrogen (Cape, 1993).
23 The possible direct effects of acidic precipitation on forest trees have been evaluated by
24 experiments on seedlings and young trees. The size of mature trees makes experimental
25 exposure difficult, therefore necessitating extrapolations from experiments on seedlings and
26 saplings; however, such extrapolations must be used with caution (Cape, 1993). Both conifers
27 and deciduous species have shown significant effects on leaf surface structures after exposure to
28 simulated acid rain or acid mist at pH 3.5. Some species have shown subtle effects at pH 4 and
29 above. Visible lesions have been observed on many species at pH 3 and on sensitive species at
30 pH 3.5 (Cape, 1993). The relative sensitivities of forest vegetation to acidic precipitation based
April 2002 4-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 on macroscopic injury have been ranked as follows: herbaceous dicots > woody dicots >
2 monocots > conifers (Percy, 1991).
3 Huttunen (1994) described the direct effects of acid rain or acidic mist on epicuticular
4 waxes whose ultrastructure is affected by plant genotype and phenotype. The effects of air
5 pollutants on epicuticular waxes of conifers have received greater study than the waxes of other
6 species. Leafage and the shorter life span of broad-leaved trees make them less indicative of the
7 effects of acid precipitation. Many experimental studies indicate that epicuticular waxes that
8 function to prevent water loss from plant leaves can be destroyed by acid rain in a few weeks
9 (Huttunen, 1994). This function is crucial in conifers because of their longevity and evergreen
10 foliage. Microscopic observations of epicuticular wax structures have, for a long time, suggested
11 links between acidic deposition and aging. In Norway spruce (Picea abies), acid rain causes not
12 only the aging of needles (which in northern conditions normally last from 11 to 14 years) to be
13 shortened, but also accelerates the erosion rate of the waxes as the needles age.
14 The effects of acidic precipitation and fog on red spruce (Picea rubens) have been studied
15 extensively (Schier and Jensen, 1992). Visible foliar injury of the needles in the form of a
16 reddish-brown discoloration has been observed on red spruce seedlings experimentally exposed
17 to acidic mist, but this visible symptom has not been observed in the field. Ultrastructural
18 changes in the epicuticular wax were observed both experimentally and on spruce growing at
19 high elevations. Laboratory studies indicate that visible injury usually does not occur unless the
20 pH is 3 or less (Schier and Jensen, 1992). Cape (1993) reported that, when compared with other
21 species, red spruce seedlings appeared to be more sensitive to acid mist. From studies of conifers
22 and a review of the literature, Huttunen (1994) concluded that acidic precipitation causes direct
23 injury to tree foliage and indirect effects through the soil. The indirect effects of acidic
24 precipitation are discussed in Section 4.3.
25 Based on a review of the many studies in the literature involving field and controlled
26 laboratory experiments on crops, Cape (1993) drew a number of conclusions concerning the
27 direct effects of acidic precipitation on crops:
28 • foliar injury and growth reduction occurs below pH 3;
29 • allocation of photosynthate is altered, with increased shoot to root ratios;
30 • expanded and recently expanded leaves are most susceptible, and injury occurs first to
31 epidermal cells;
April 2002 4-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 • leaf surface characteristics such as wettability, buffering capacity, and transport of
2 material across the leaf surface contribute to susceptibility and differ among species;
3 • data obtained from experiments in greenhouses or controlled environmental chambers
4 cannot be used to predict effects on plants grown in the field;
5 • quantitative data from experimental exposures cannot be extrapolated to field exposures
6 because of differences and fluctuations in concentrations, durations, and frequency of
7 exposure;
8 • there are large differences in response within species;
9 • timing of exposure in relation to phenology is of utmost importance;
10 • plants may be able to recover from or adapt to injurious exposures; and
11 • sequential exposure to acidic precipitation and gaseous pollutants is unlikely to be more
12 injurious than exposure to individual pollutants.
13 Studies by Chevone et al. (1986), Krupa and Legge (1986), and Blaschke (1990) differ with
14 the last conclusion of Cape listed above. Their studies indicate that interactions between acidic
15 deposition and gaseous pollutants do occur. Acidity affects plant responses to both O3 and SO2.
16 Chevone et al. (1986) observed increased visible injury on soybean and pinto bean when acid
17 aerosol exposure preceded O3 exposure; whereas linear decreases in dry root weight of yellow
18 poplar occurred as acidity increased with simultaneous exposures to O3 and simulated acid rain.
19 Krupa and Legge (1986) also noted increased visible injury to pinto bean plants when aerosol
20 exposure preceded O3 exposure. In none of the studies cited above did acid rain per se produce
21 significant growth changes. In contrast, Blaschke (1990) observed a decrease in ectomycorrhizal
22 frequency and short root distribution caused by acid rain exposure in combination with either
23 SO2 or O3.
24
25 Trace Elements. All but 10 of the 90 elements that comprise the inorganic fraction of the
26 soil occur at concentrations of less than 0.1% (1000 Mg/g) and are termed "trace" elements.
27 Trace elements with a density greater than 6 g cm"3, referred to as "heavy metals," are of
28 particular interest because of their potential toxicity for plant and animals. Although some trace
29 metals are essential for vegetative growth or animal health, they are all toxic in large quantities.
30 Combustion processes produce metal chlorides that tend to be volatile and metal oxides that tend
31 to be nonvolatile in the vapor phase (McGowan et al., 1993). Most trace elements exist in the
April 2002 4-58 DRAFT-DO NOT QUOTE OR CITE
-------
1 atmosphere in particulate form as metal oxides (Ormrod, 1984). Aerosols containing trace
2 elements derive predominantly from industrial activities (Ormrod, 1984). Generally, only
3 cadmium, chromium, nickel, and mercury are released from stacks in the vapor phase (McGowan
4 etal., 1993). Concentrations of heavy metals in incinerator fly ash increase with decreasing
5 particle size.
6 Vegetational surfaces, especially the foliage, present a major reaction and filtration surface
7 to the atmosphere and act to accumulate particles deposited via wet and dry processes described
8 in Section 4.2.1 (long, 1991; Youngs et al., 1993). The chemical constituents of particles
9 deposited on foliar surfaces may be taken up through the leaf surface. The greatest particle
10 loading is usually on the adaxial (upper) leaf surface where particles accumulate in the mid-vein,
11 center portion of the leaves. Additionally, the mycelium of fungi becomes particularly abundant
12 on leaf surfaces as the growing season progresses and is in intimate association with deposited
13 particles (Smith, 1990c).
14 Investigations of trace elements present along roadsides and in industrial and urban
15 environments indicate that impressive burdens of parti culate heavy metals can accumulate on
16 vegetative surfaces. Foliar uptake of available metals could result in metabolic effects in above-
17 ground tissues. Only a few metals, however, have been documented to cause direct phytotoxicity
18 in field conditions. Copper, zinc, and nickel toxicities have been observed most frequently. Low
19 solubility, however, limits foliar uptake and direct heavy metal toxicity because trace metals
20 must be brought into solution before they can enter into leaves or bark of vascular plants.
21 In those instances when trace metals are absorbed, they are frequently bound in leaf tissue and
22 are lost when the leaf drops off (Hughes, 1981). Trace metals in mixtures may interact to cause a
23 different plant response when compared with a single element; however, there has been little
24 research on this aspect (Ormrod, 1984). In experiments using chambers, Marchwinska and
25 Kucharski (1987) studied the effects of SO2 alone and in combination with PM components (Pb,
26 Cd, Zn, Fe, Cu, and Mn) obtained from a zinc smelter bag filter. The combined effects of SO2
27 and PM further increased the reduction in yield of beans caused by SO2; whereas the
28 combination, though severely injuring the foliage, produced little effect on carrots and parsley
29 roots except after long-term exposures (when there was a decrease in root weight).
30 Trace metal toxicity of lichens has been demonstrated in relatively few cases. Nash (1975)
31 documented zinc toxicity in the vicinity of a zinc smelter near Palmerton, PA. Lichen species
April 2002 4-59 DRAFT-DO NOT QUOTE OR CITE
-------
1 richness and abundance were reduced by approximately 90% in lichen communities at Lehigh
2 Water Gap near the zinc smelter when compared with those at Delaware Water Gap. Zinc,
3 cadmium, and sulfur dioxide were present in concentrations toxic to some species near the
4 smelter; however, toxic zinc concentrations were detected farther away than the detectable limits
5 of sulfur dioxide (Nash, 1975). Experimental data suggest that lichen tolerance to Zn and Cd
6 falls between 200 and 600 ppm (Nash, 1975).
7 Though there has been no direct evidence of a physiological association between tree injury
8 and exposure to metals, heavy metals have been implicated because their deposition pattern is
9 correlated with forest decline. The role of heavy metals has been indicated by phytochelatin
10 measurements. Phytochelatins are intracellular metal-binding peptides that act as specific
11 indicators of metal stress. Because they are produced by plants as a response to sublethal
12 concentrations of heavy metals, they can indicate that heavy metals play a role in forest decline
13 (Gawel et al., 1996). Concentrations of heavy metals increased with altitude, as did forest
14 decline, and increased concentrations across the study region that show increased levels of forest
15 injury, as well.
16 Phytochelatin concentrations were measured in red spruce and balsam fir (Abies balsamea)
17 needles throughout the 1993 growing season at 1000 m on Whiteface Mountain in New York.
18 Mean foliar concentrations in red spruce were consistently higher than in balsam fir from June
19 until August, with the greatest and most significant difference occurring at the peak of the
20 growing season in mid-July. In July, the phytochelatin concentrations were significantly higher
21 than at any other time measured. Balsam fir did exhibit this peak, but maintained a consistently
22 low level throughout the season. Both the number of dead red spruce trees and phytochelatin
23 concentrations increased sharply with elevation (Gawel et al., 1996). The relationship between
24 heavy metals and the decline of forests in northeastern United States was further tested by
25 sampling red spruce stands showing varying degrees of decline at 1000 m on nine mountains
26 spanning New Hampshire, Vermont, and New York. The collected samples indicated a
27 systematic and significant increase in phytochelatin concentrations associated with the extent of
28 tree injury. The highest phytochelatin concentrations were measured during 1994 from sites
29 most severely affected by forest decline in the Green Mountains, VT, and the Adirondack
30 Mountains, NY. These data strongly imply that metal stress causes tree injury and contributes to
31 forest decline in the northeastern United States (Gawel et al., 1996).
April 2002 4-60 DRAFT-DO NOT QUOTE OR CITE
-------
1 One potential direct effect of heavy metals is on the activity of microorganisms and
2 arthropods resident on and in the leaf surface ecosystem. The fungi and bacteria living on and in
3 the surfaces of leaves play an important role in the microbial succession that prepares leaves for
4 decay and litter decomposition after their fall (U.S. Environmental Protection Agency, 1996b).
5 Numerous fungi were consistently isolated from foliar surfaces at various crown positions
6 from London plane trees growing in roadside environments in New Haven, CT. Those existing
7 primarily as saprophytes included Aureobasidium pullulans, Chaetomium sp., Cladosporium sp.,
8 Epicoccum sp., and Philaphora verrucosa. Those existing primarily as parasites included
9 Gnomoniaplatani, Pestalotioposis sp., and Pleurophomella sp. The following cations were
10 tested in vitro for their ability to influence the growth of these fungi: cadmium, copper,
11 manganese, aluminum, chromium, nickel, iron, lead, sodium, and zinc. Results indicated
12 variable fungal response with no correlation between saprophytic or parasitic activity and
13 sensitivity to heavy metals. Both linear extension and dry weight data indicated that the
14 saprophytic Chaetomum sp. was very sensitive to numerous metals. Aureobasidium pullulans,
15 Epicoccum sp.., and especially P.verrucosa, on the other hand, appeared to be much more
16 tolerant. Of the parasites, G. platani appeared to be more tolerant than Pestalotiopsis sp. and
17 Pleurophomella sp. Metals exhibiting the broadest spectrum growth suppression were iron,
18 aluminum, nickel, zinc, manganese, and lead (Smith and Staskawicz, 1977; Smith, 1990c).
19 These in vitro studies employed soluble compounds containing heavy metals. Trace metals
20 probably occur naturally on leaf surfaces as low-solubility oxides, halides, sulfates, sulfides, or
21 phosphates (Clevenger et al., 1991; Koslow et al., 1977). In the event of sufficient solubility and
22 dose, however, changes in microbial community structure on leaf surfaces because of heavy
23 metal accumulation are possible.
24
25 Organic Compounds. Fine particles in the atmosphere reacting with volatilized chemical
26 compounds are partitioned between the gas and particle phases, depending on the liquid phase
27 vapor pressure at the ambient atmospheric temperature, the surface area of the particles per unit
28 volume of air, the nature of the particles and of the chemical being adsorbed; and they can be
29 removed by wet and dry deposition (McLachlan, 1996a). Materials as diverse as DDT,
30 polychlorinated biphenyls (PCBs), and polynuclear aromatic hydrocarbons (PAHs) are being
31 deposited from the atmosphere on rural as well as urban landscapes (Kylin et al., 1994). Motor
April 2002 4-61 DRAFT-DO NOT QUOTE OR CITE
-------
1 vehicles emit particles to the atmosphere from several sources in addition to the tailpipe. Rogge
2 et al. (1993b) inventoried the organic contaminants associated with fine particles (diameter
3 <2.0 //m) in road dust, brake-lining-wear particles, and tire tread debris. More than 100 organic
4 compounds were identified in these samples, including n-alkanols, benzoic acids, benzaldehydes,
5 polyalkylene glycol ethers, PAHs, oxy-PAH, steranes, hopanes, natural resins, and other
6 compound classes. A large number of PAHs, ranging from naphthalene (C10H8) to 5- and 6-ring
7 and higher PAHs, their alkyl-substituted analogues, and their oxygen- and nitrogen-containing
8 derivatives are emitted from motor vehicle sources (Seinfeld, 1989).
9 Plants may be used as environmental monitors to compare the deposition of PAH,
10 persistent organic pollutants (POPs), or semivolatile organic components (SOCs) between sites
11 (e.g., urban versus rural; Wagrowski and Kites, 1997; Ockenden et al., 1998; McLachlan, 1999).
12 Vegetation can be used qualitatively to indicate organic pollutant levels as long as the mechanism
13 of accumulation is considered. The substance may enter the plant via the roots or, as mentioned
14 above, be deposited as a particle onto the waxy cuticle of leaves or be taken up through the
15 stomata. The pathways are a function of the chemical and physical properties of the pollutant
16 such as its lipophilicity, water solubility, vapor pressure (which controls the vapro-particle
17 partitioning) and Henry's law constant; environmental conditions such as ambient temperature
18 and the organic content of the soil; and the plant species, which controls the surface area and
19 lipids available for accumulation (Simonich and Kites, 1995). Ockenden et al. (1998) have
20 observed that, for lipophilic POPs, atmospheric transfer to plant has been the main avenue of
21 accumulation. Plants can differentially accumulate POPs. Results have shown differences
22 between species with higher concentrations in the lichen (Hypogymniaphysiodes) than in Scots
23 pine needles (Pinus sylvestris). Even plants of the same species, because they have different
24 growth rates and different lipid contents (depending on the habitat in which they are growing),
25 have different rates of sequestering pollutants. These facts confound data interpretations and
26 must be taken into account when considering their use as passive samplers.
27 Vegetation itself is an important source of hydrocarbon aerosols. Terpenes, particularly
28 a-pinene, p-pinene, and limonene, released from tree foliage may react in the atmosphere to form
29 submicron particles. These naturally generated organic particles contribute significantly to the
30 blue haze aerosols formed naturally over forested areas (Geron et al., 2000).
April 2002 4-62 DRAFT-DO NOT QUOTE OR CITE
-------
1 The low water solubility with high lipoaffinity of many of these organic xenobiotics
2 strongly control their interaction with the vegetative components of natural ecosystems. The
3 cuticles of foliar surfaces are covered with a wax layer that helps protect plants from moisture
4 and short-wave radiation stress. This epicuticular wax, consisting mainly of long-chain esters,
5 polyesters, and paraffins, has been demonstrated to accumulate lipophilic compounds. Organic
6 air contaminants in the particulate or vapor phase are absorbed to and accumulate in the
7 epicuticular wax of vegetative surfaces (Gaggi et al., 1985; Kylin et al., 1994). Direct uptake of
8 organic contaminants through the cuticle or the vapor-phase uptake through the stomates are
9 characterized poorly for most trace organics. The phytotoxicity and toxicity of organic
10 contaminants to soil microorganisms is not well studied (Foster, 1991).
11
12 Summary
13 Particulate matter transferred from the atmosphere may be deposited on above-ground plant
14 parts and may exert physical or chemical effects or both. The effects of dust deposited on plant
15 surfaces are more likely to be associated with their chemistry than simply with the mass of
16 deposited particles. Studies of the effects of chemicals in PM deposited on foliage have found
17 little or no effects on foliar processes unless exposure levels were significantly higher than
18 typically would be experienced in the ambient environment. The majority of easily identified
19 direct effects, other than climate, occur in severely polluted areas around heavily industrialized
20 point sources such as limestone quarries, cement kilns, and smelting facilities for iron, lead, or
21 various other metals. The direct effects of PM on foliar surfaces are confounded by the chemical
22 nature and size characteristics of ambient airborne particles and the absence of a clear distinction
23 between effects of PM on foliar surfaces and effects attributed to forms of air pollutants. Most
24 documented toxic effects of particles on vegetation are associated with their acidity, trace metal
25 content, nutrient content, surfactant properties, or salinity.
26
27 4.2.2.2 Indirect Effects of Particulate Matter on Natural Ecosystems
28 All life on this planet is dependent on the chemical energy in the form of carbon
29 compounds to sustain their life processes. Terrestrial vegetation, via the process of
30 photosynthesis, provides approximately half of the carbon that annually cycles between the Earth
31 and the atmosphere (Chapin and Ruess, 2001). Plants do not live alone. They are members of
April 2002 4-63 DRAFT-DO NOT QUOTE OR CITE
-------
1 ecosystems, structurally complex communities comprised of populations of plant, animals
2 (including humans), insects, and microorganisms that interact with one another and with their
3 non-living (abiotic) chemical and physical environment. Ecosystems are dynamic, self-adjusting,
4 self-maintaining, complex and adaptive systems, in which patterns at the higher levels of
5 organization emerge from interactions and selection processes at localized levels (Levin, 1998).
6 Ecosystem components must have an adequate supply of energy, mineral nutrients, and
7 water to maintain themselves and function properly. Sunlight is the energy source for most
8 ecosystems. The energy obtained by plants (the producers) from sunlight during photosynthesis
9 and chemical nutrients (e.g., nitrogen, phosphorus, sulfur) taken up from the soil are transferred
10 to other species (the consumers) within the ecosystem through food webs. The movement of
11 chemical nutrients through an ecosystem is cyclic, as the nutrients are used or stored and
12 eventually returned to the soil by decomposer organisms. Energy, on the other hand, is
13 transferred from organism to organism through an ecosystem in food webs and, finally, is
14 dissipated into the atmosphere as heat (Odum, 1993). The flows of energy and cycling of
15 nutrients provide the interconnectedness between ecosystem parts and transforms the community
16 from a random collection of species into an integrated whole, an ecosystem, in which the biotic
17 and abiotic parts are interrelated (Levin, 1998).
18 Macroscopic ecosystem properties (such as structure, diversity-productivity relationships,
19 and patterns of nutrient and energy movement) that emerge from the interactions among the
20 various components may feed back to influence subsequent development of those interactions.
21 The relationship between structure and function is a fundamental one in ecosystem science.
22 Ecosystem structure refers to the component species, their biodiversity, abundance, mass, and
23 arrangement within an ecosystem. Ecosystem functions (energy flow, nutrient flux, water and
24 material flow) are characterized by the way in which the components (e.g., plants, animals, and
25 microorganisms) interact and the effect their activities have on the physical and chemical
26 environment. Elucidating these interactions across scales is fundamental to understanding the
27 relationships between biodiversity and ecosystem functioning (Levin, 1998).
28 Both ecosystem structure and functions play an essential role in providing ecosystem
29 services. Human existence on this planet depends on the life-support services provided by
30 ecosystem structure and functions (Daily, 1997). Ecosystem functions are characterized by the
31 way components interact. These are the functions that maintain clean water, pure air, a vegetated
April 2002 4-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 earth, and a balance of organisms, the functions that enable humans to survive. They are the
2 dynamics of ecosystems. The benefits they impart include absorption and breakdown of
3 pollutants, cycling of nutrients, binding of soil, degradation of organic waste, maintenance of a
4 balance of gases in the air, regulation of radiation balance, climate, and the fixation of solar
5 energy (Table 4-13; Westman, 1977; Daily, 1997). Economic benefits and values associated
6 with ecosystem functions and services, and the need to preserve them because of their value to
7 human life, are discussed by Costanza et al. (1997) and (Pimentel et al., 1997). Services usually
8 are not considered to be items with market value.
9
10
TABLE 4-13. ECOSYSTEM SERVICES
• Purification of air and water - functions
• Mitigation of floods and droughts - structure and functions
• Detoxification and decomposition of wastes - functions
• Generation and renewal of soil and soil fertility - functions
• Pollination of crops and natural vegetation - functions
• Control of the vast majority of potential agricultural pests - functions
• Dispersal of seeds and translocation of nutrients - functions
• Maintenance of biodiversity, from which humanity has derived key elements of its
agricultural, medicinal, and industrial enterprises
• Partial stabilization of climate
• Moderation of temperature extremes and the force of winds and waves
• Support of diverse human cultures
• Providing of aesthetic beauty and intellectual stimulation that lift the human spirit
Source: Daily (1997).
1 Concern has risen in recent years regarding the consequences of changing the biological
2 diversity of ecosystems (Tilman, 2000; Ayensu et al., 1999; Wall, 1999; Hooper and Vitousek,
3 1997; Chapin et al., 1998). The concerns arise because human activities are creating
4 disturbances that are causing the loss of biodiversity, altering the complexity and stability of
5 ecosystems, and producing changes in nutrient cycling and the structure and function of
6 ecosystems (Pimm, 1984; Levin, 1998; Chapin et al., 1998; Peterson et al., 1998; Tilman, 1996;
April 2002 4-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 Tilman and Downing, 1994; Wall, 1999; Daily and Ehrlich, 1999). Changes in ecosystem
2 structure and function affect the ecosystem services vital to human life.
3 There are few ecosystems on earth today that are not influenced by humans (Freudenburg
4 and Alario, 1999; Vitousek et al., 1997; Matson et al. 1997; Noble and Dirzo, 1997). The
5 scientific literature is filled with publications discussing the importance of ecosystem structure
6 and function. Eco-risk, complexity, stability, biodiversity, resilience, sustainability,
7 management, risk assessment, and ecosystem health, are frequently discussed topics. The
8 deposition of particulate matter from the atmosphere has the potential to alter ecosystem structure
9 and function by altering nutrient cycling and changing biodiversity. There is a need, therefore, to
10 understand how ecosystems respond to both natural and anthropogenic stresses and, especially,
11 the ways that anthropogenic stresses are impacting ecosystem services and products.
12 Specifically, understanding the ecological effects of PM deposition is as important as quantifying
13 the human health effects of PM deposition.
14
15 Ecosystem Response to Stresses
16 Ecosystem response to stress begins at the population level. Population change, however,
17 begins with the response of individual plants or animals. Plant responses, both structural and
18 functional, must be scaled in both time and space and propagated from the individual to the more
19 complex levels of community interaction to produce observable changes in an ecosystem (see
20 Figure 4-7). In an ecosystem, at least three levels of biological interaction are involved: (l)the
21 individual plant and its environment; (2) the population and its environment; and (3) the
22 biological community composed of many species and its environment (Billings, 1978).
23 Individual organisms within a population vary in their ability to withstand the stress of
24 environmental change. The response of individual organisms within a population is based on
25 their genetic constitution (genotype), stage of growth at time of exposure to stress, and the
26 microhabitat in which they are growing (Levin, 1998). The range within which these organisms
27 can exist and function determines the ability of the population to survive. Those able to cope
28 with the stresses survive and reproduce. Competition among the different species results in
29 succession (community change over time) and, ultimately, produces ecosystems composed of
30 populations of plant species that have the capability to tolerate the stresses (Rapport and
31 Whitford, 1999; Guderian, 1985).
April 2002 4-66 DRAFT-DO NOT QUOTE OR CITE
-------
N. Reactic
_evel ot
Organization
Leaf
(cm2)
Branch
(cm2)
Tree
(nf)
Stand
(ha)
n Time
Minute
Day
Year
»•
Ni •
te
mx_ |
X-^J
•^
•
Decade
1
3
•-4
^5
^
\^J
1
Century
•10
11
-12
-1 7
s\
16
Injury Symptom
Needle necrosis
and abscission
Branch length,
bifurcation ratio,
and ring-width
growth altered
Reduction in
diameter and death
of tree
Decreases in
stand productivity,
increases in mortality
and alterations in
regeneration patterns
Key Changes in Processes
Reduced carbon assimilation
because of reduced radiation
Reduced carbon available for foliage
replacement and branch growth/
export Synergistic interaction
between mistletoe and tephra
deposition
Reduced carbon available for
height, crown, and stem growth
Influence of crown class on initial
impact and subsequent recovery
Interaction between stand
composition and recovery
For a given level, the dot associated with a line begins with a process (e.g. photosynthesis for#1 under leaf)
and ends with the associated structure (e.g., the needle).
Evaluating Impacts Within a Level of Organization
Leaf Level Carbon exchange-1
Carbon pools-2
Needle number and size-3
Needle retention/abscission-4
Branch Level Carbon allocation-5
Branch growth-6
Branch morphology-7
Branch vigor-8
Branch retention-9
Tree Level Height and diameter growth-10
Crown shape and size-11
Tree vigor-12
Mortality-13
Stand Level Productivity-14
Mortality-15
Species composition-16
Evaluating Interactions Between Different Levels of Organization
The diagonal arrow indicates the interaction between any two levels of organizdion.
The types of interaction are due to the properties of variability and compensation.
A- Refers to the interaction between the leaf and branch levels, where, for example,
variability at the branch level determines leaf quantity, and compensation at the leaf
level in photosynthesis may compensate for the reduction in foliage amount.
B - Refers to the interaction between the branch and the tree, where variability in branches
determines initial interception, branch vigor, and branch location in the crown;
compensation may be related to increased radiation reaching lower branches.
C - Refers to the interaction between the tree and the stand. Both genetic and
environmental variability, inter- and intraspecific compensations, and tree historical
and competitive synergisms are involved.
Figure 4-7. Effects of environmental stress on forest trees are presented on a hierarchial
scale for the leaf, branch, tree, and stand levels of organization. The
evaluation of effects within a level of organization are indicated by horizontal
arrows. The evaluation of interactions between different levels of organization
are indicated by diagonal arrows.
Source: Hinckley et al. (1992).
April 2002
4-67
DRAFT-DO NOT QUOTE OR CITE
-------
1 The number of species in a community usually increases during succession in unpolluted
2 atmospheres. Productivity, biomass, community height, and structural complexity increase.
3 Severe stresses, on the other hand, divert energy from growth and reproduction to maintenance
4 and return succession to an earlier stage (Waring and Schlesinger, 1985). Ecosystems are subject
5 to natural periodic stresses, such as drought, flooding, fire, and attacks by biotic pathogens (e.g.,
6 fungi, insects). Extremely severe natural perturbations return succession to an earlier stage;
7 reduce ecosystem structure and functions (i.e., produce a scarcity of life forms and extinguish
8 symbiotic interactions); disrupt the plant processes of photosynthesis and nutrient uptake, carbon
9 allocation, and transformation that are directly related to energy flow and nutrient cycling;
10 shorten food chains; and reduce the total nutrient inventory (Odum, 1993). This transformation,
11 however, sets the stage for recovery that permits the perturbed ecosystem to adapt to changing
12 environments (Holling, 1986). Therefore, these perturbations are seldom more than a temporary
13 setback, and recovery can be rapid (Odum, 1969).
14 In contrast, anthropogenic stresses usually are severe, debilitating stresses. Severely
15 stressed ecosystems do not recover readily, but may be further degraded (Odum, 1969; Rapport
16 and Whitford, 1999). Anthropogenic stresses can be classified into four main groups:
17 (1) physical restructuring (e.g., changes resulting from land use); (2) introduction of exotic
18 species; (3) over harvesting; and (4) discharge of toxic substances into the atmosphere, onto land,
19 and into water. Ecosystems lack the capacity to adapt to the above stresses and maintain their
20 normal structure and functions unless the stress is removed (Rapport and Whitford, 1999). These
21 stresses result in a process of degradation marked by a decrease in biodiversity, reduced primary
22 and secondary production, and a lower capacity to recover and return to its original state.
23 In addition, there is an increased prevalence of disease, reduced nutrient cycling, increased
24 dominance of exotic species, and increased dominance by smaller, short-lived opportunistic
25 species (Odum, 1985; Rapport and Whitford, 1999). Discharge of toxic substances into the
26 atmosphere, onto land, and into water can cause acute and chronic stresses; and, once the stress is
27 removed, a process of succession begins that can ultimately return the ecosystem to a semblance
28 of its former structure. Air pollution stresses, if acute, are usually short term and the effects soon
29 visible. Chronic stresses, on the other hand, are long-term stresses whose effects occur at
30 different levels of ecosystem organization and appear only after long-term exposures, as in the
April 2002 4-68 DRAFT-DO NOT QUOTE OR CITE
-------
1 case of acidic deposition in the northeast or ozone in California (Shortle and Bondietti, 1992;
2 U.S. Environmental Protection Agency, 1996b).
3 The possible effects of air pollutants on ecosystems have been categorized by Guderian
4 (1977) as follows:
5
6 (1) accumulation of pollutants in the plant and other ecosystem components (such as soil
7 and surface- and groundwater),
8 (2) damage to consumers as a result of pollutant accumulation,
9 (3) changes in species diversity because of shifts in competition,
10 (4) disruption of biogeochemical cycles,
11 (5) disruption of stability and reduction in the ability of self-regulation,
12 (6) breakdown of stands and associations, and
13 (7) expansion of denuded zones.
14
15 How changes in these functions can result from PM deposition and influence ecosystems is
16 discussed in the following text. It should be remembered that, although the effects of PM are
17 being emphasized, the vegetational components of ecosystems also are responding to multiple
18 stresses from other sources.
19
20 Ecosystem Response to Direct Plant Effects
21 The presence of PM in the atmosphere may affect vegetation directly, following physical
22 contact with the foliar surface (Section 4.2), but in most cases, the more significant effects are
23 indirect. These effects may be mediated by suspended PM (i.e., through effects on radiation and
24 climate) and by particles that pass through the vegetative canopies to the soil. Paniculate matter,
25 as considered in this chapter is a heterogeneous mixture of particles differing in size, origin, and
26 chemical constituents, and their effects vary depending on the chemical nature of PM being
27 deposited on vegetation or soil. Particulate inputs and ecosystem cycling of key elements are
28 considered below.
29 The majority of studies dealing with direct effects of particulate dust and trace metals on
30 vegetation have focused on responses of individual plant species and were conducted in the
31 laboratory or in controlled environments (Saunders and Godzik, 1986). A few have considered
April 2002 4-69 DRAFT-DO NOT QUOTE OR CITE
-------
1 the effects of particles on populations, communities, and ecosystems. Most of these focused on
2 ecosystems in industrialized areas heavily polluted by deposits of both chemically inert and
3 active dusts. Effects can result from direct deposition or indirectly by deposition onto the soil.
4 Reductions in growth, yield, flowering, and reproduction of plants from particulate deposition
5 have been reported (Saunders and Godzik, 1986). Sensitivities of individual species have been
6 associated with changes in composition and structure of natural ecosystems.
7 Evidence from studies of effects of PM deposition, specifically chemically inert and active
8 dusts indicates that, within a population, plants exhibit a wide range of sensitivity, which is the
9 basis for the natural selection of tolerant individuals. Rapid evolution of certain populations of
10 tolerant species at sites with heavy trace element and nitrate deposition has been observed
11 (Saunders and Godzik, 1986). Tolerant individuals present in low frequencies in populations
12 when growing in unpolluted areas have been selected for tolerance at both the seedling and adult
13 stages when exposed to trace metal or nitrate deposition (Ormrod, 1984; U.S. Environmental
14 Protection Agency, 1993). Chronic pollutant injury to a forest community may result in the loss
15 of sensitive species, loss of tree canopy, and maintenance of a residual cover of pollutant-tolerant
16 herbs or shrubs that are recognized as successional species (Table 4-14; Smith, 1974).
17 Responses of ecosystems to stresses (unless severe or catastrophic) are difficult to
18 determine because the changes are subtle (Garner, 1991). This is particularly true of responses to
19 particles. Changes in the soil may not be observed until accumulation of the pollutant has
20 occurred for 10 or more years, except in the severely polluted areas around heavily industrialized
21 point sources (Saunders and Godzik, 1986). In addition, the presence of other co-occurring
22 pollutants makes it difficult to attribute the effects to PM alone. In other words, the potential for
23 alteration of ecosystem function and structure exists but is difficult to quantify, especially when
24 there are other pollutants present in the ambient air that may produce additive or synergistic
25 responses even though PM concentrations may not be elevated.
26
27 Physical Effects
28 The direct effects of limestone dust on plants and ecosystems has been known for many
29 years. Long-term changes in the structure and composition of the seedling-shrub and sapling
30 strata of an experimental site near limestone quarries and processing plants in Giles County in
31 southwestern Virginia were reported by Brandt and Rhoades (1972, 1973). Dominant trees in the
April 2002 4-70 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-14. ECOSYSTEM FUNCTIONS IMPACTED BY AIR POLLUTION
EFFECTS ON TEMPERATE FOREST ECOSYSTEMS
Forest Soil and Vegetation: Activity and Response
Ecosystem Consequence and Impact
1. Forest tree reproduction, alteration, or inhibition
2. Forest nutrient cycling, alteration
a. Reduced litter decomposition
b. Increased plant and soil leaching and soil
weathering
c. Disturbance of microbial symbioses
3. Forest metabolism
a. Decreased photosynthesis
b. Increased respiration
c. Altered carbon allocation
4. Forest stress, alteration
a. Phytophagous insects, increased or decreased
activity
b. Microbial pathogens, increased or decreased
activity
c. Foliar damage increased by direct air pollution
influence
1. Altered species composition
2. Reduced growth, less biomass
3. Reduced growth, less biomass
4. Altered ecosystem stress:
increased or decreased insect infestations;
increased or decreased disease epidemics;
and reduced growth, less biomass, and
altered species composition
Source: Smith (1974).
1 control area, a part of the oak-chestnut association of the eastern deciduous forests of eastern
2 North America, were chestnut oak (Quercusprinus), red oak (Q. rubra), and red maple
3 (Acer rubrum). An abundance of uniformly distributed saplings and seedlings were visible under
4 the tree canopy, and herbs appeared in localized areas in canopy openings. Chestnut oak
5 dominated the area, and the larger trees were 60 to 80 years old. The dusty site was dominated
6 by white oak (Q. alba); whereas red oak and tulip poplar (Liriodendron tulipifera) were
7 subcodominants. The largest trees were 100 years old and had necrotic leaves, peeling bark, and
8 appeared to be in generally poor condition except for tulip poplar (which thrived in localized
9 areas). The site contained a tangled growth of seedlings and shrubs, a few saplings, and a
10 prevalence of green briar (Smilax spp.) and grape (Vitis spp). The sapling strata in the area was
11 represented by red maple, hickory (Carya spp.), dogwood (Cornus florida), and hop-hornbeam
12 (Ostrya virginiana). Saplings of none of the leading dominant trees were of importance in this
13 stratum. The most obvious form of vegetation in the seedling-shrub stratum, because of their
April 2002
4-71
DRAFT-DO NOT QUOTE OR CITE
-------
1 tangled appearance, were dogwood, hop-hornbeam, redbud (Cercis canadensis), and sugar maple
2 (Acer saccarum).
3 Crust formation reduced photosynthesis, induced premature leaf fall and destruction of leaf
4 tissues, inhibited growth of new tissue, and reduced the formation of carbohydrate needed for
5 normal growth and storage (Brandt and Rhoades, 1973). The authors (Brandt and Rhoades,
6 1972), citing Odum (1969), also stated that one result of the accumulation of toxic pollutants in
7 the biosphere (as the result of human activities) is the simplification of both plant and animal
8 communities. In plant communities, structure is determined by sampling various strata within
9 the community. Each stratum comprises a particular life form (e.g., herbs, seedlings, saplings,
10 trees). Dust accumulation favored growth of some species and limited others. For example,
11 sugar maple was more abundant in all strata of the dusty site when compared with the control site
12 where it was present only as a seedling. The growth of tulip poplar, dogwood, hop-hornbeam,
13 black haw (Viburnum prunifolium), and redbud (C. canadensis) appeared to be favored by the
14 dust. Growth of conifers and acidophiles such as rhododendron (Rhododendron maximum),
15 however, was limited. Although dust accumulation began in 1945, the heaviest accumulation
16 occurred between 1967 and 1972 during the time of the study.
17 Changes in community composition were associated closely with changes in the growth of
18 the dominant trees. Decrease in density of seedlings and saplings and in mean basal area, as well
19 as lateral growth of red maple, chestnut oak, and red oak, occurred in all strata. On the other
20 hand, all of these characteristics increased in tulip poplar, which was a subordinate species before
21 dust accumulation began but had assumed dominance at the time of the study. Reduction in
22 growth of the dominant trees had apparently given tulip poplar competitive advantage because of
23 its ability to tolerate dust. Changes in soil alkalinity occurred because of the heavy deposition of
24 limestone dust; however, the facilities necessary for critical analysis of the soils were not
25 available. From the foregoing, it is obvious that PM physical effects in the vicinity of limestone
26 quarries and processing plants can affect ecosystems.
27 Changes in ecosystem structure resulting from exposures to sea salt were cited previously
28 (Section 4.3.1.1). The dominance of live oak (Quercus virginiand) as a practically pure stand on
29 Smith Island (Bald Head), NC and along the eastern and southern coast of North Carolina has
30 been explained as due to its tolerance to salt spray. The absence of more inland species is
April 2002 4-72 DRAFT-DO NOT QUOTE OR CITE
-------
1 attributed to their intolerance to salt spray. Wells (1939) termed the long-term stabilization of
2 live oak as "salt spray climax," a new type of climax.
3
4 Acidic Deposition
5 The effects of acidic deposition have been discussed in several previous reports. The 1982
6 EPA document, Air Quality Criteria for Particulate Matter and Sulfur Oxides, devoted a chapter
7 to the effects of acidic deposition (U.S. Environmental Protection Agency, 1982). In 1984, EPA
8 published The Acidic Deposition Phenomenon and Its Effects (Altshuller and Linthurst, 1984),
9 and, in 1991, NAPAP published the result of its extensive study, Acidic Deposition: State of
10 Science and Technology (Irving, 1991). The major effects of acidic deposition occur through the
11 soil and are discussed under indirect effects. However, included among the direct responses of
12 forest trees to acidic deposition are increased leaching of nutrients from foliage; accelerated
13 weathering of leaf cuticular surfaces; increased permeability of leaf surfaces to toxic materials,
14 water, and disease agents; and altered reproductive processes (Altshuller and Linthurst, 1984).
15
16 Trace Elements
17 Possible direct responses of trace elements on vegetation result from their deposition and
18 residence on the phyllosphere (i.e., leaf surfaces). Fungi and other microorganisms living on the
19 leaves of trees and other vegetation play an important role in leaf decomposition after litterfall
20 (Miller and McBride, 1999; Jensen, 1974; Millar, 1974). Possible impacts of heavy metals on
21 nutrient cycling and their effects on leaf microflora appear not to have been studied.
22 A trace metal must be brought into solution before it can enter into the leaves or bark of
23 vascular plants. Low solubility limits entry. In those instances when trace metals are absorbed,
24 they frequently are bound in the leaf tissue and then are lost from the plant when the leaf drops
25 off (Hughes, 1981) are transferred to the litter layer where they can affect litter decomposition, an
26 important source of soil nutrients. Changes in litter decomposition processes influence nutrient
27 cycling in the soil and limit the supply of essential nutrients. Both Cotrufo et al. (1995) and
28 Niklinska et al. (1998) point out that heavy metals affect forest litter decomposition. Cotrufo
29 et al. (1995) observed that decomposition of oak leaves containing Fe, Zn, Cu, Cr, Ni, and Pb
30 was influenced strongly during the early stages by metal contamination. Fungal mycelium was
31 significantly less abundant in litter and soil in contaminated sites when compared with control
April 2002 4-73 DRAFT-DO NOT QUOTE OR CITE
-------
1 sites. Niklinska et al. (1998) stated that toxic effects of heavy metals on soil respiration rate have
2 been reported by many scientists, and that, in polluted environments, this results in accumulation
3 of undecomposed organic matter. However, they state that results of experiments should identify
4 the most important "natural" factors affecting soil/litter sensitivity because the effects of heavy
5 metals on respiration rates depend on the dose of heavy metals, the type of litter, types of metals
6 deposited, and the storage time before respiration tests are made.
7 Trace metals, particularly heavy metals (e.g., cadmium, copper, lead, chromium, mercury,
8 nickel, zinc) have the greatest potential for influencing forest growth (Smith, 1991).
9 Experimental data indicate that the broadest spectrum of growth suppression of foliar microflora
10 resulted from iron, aluminum, and zinc. These three metals also inhibited spore formation, as did
11 cadmium, chromium, manganese, and nickel (see Smith, 1990c). In the field, the greatest injury
12 occurs from pollution near mining, smelting, and other industrial sources (Ormrod, 1984). Direct
13 metal phytotoxicity can occur only if the metal can move from the surface into the leaf or directly
14 from the soil into the root.
15
16 Organic Compounds
17 Secondary organic compounds formed in the atmosphere, the effects of some of which are
18 discussed below, have been referred to under the following terms: toxic substances, pesticides,
19 hazardous air pollutants (HAPS), air toxics, semivolatile organic compounds (SOCs), and
20 persistent organic pollutants (POPS). Again, it should be noted that the chemical substances
21 denoted by such headings are not criteria air pollutants controlled by the NAAQS under
22 Section 109 of the Clean Air Act (U.S. Code, 1991), but rather are controlled under Sect. 112,
23 Hazardous Air Pollutants. Their possible effects on humans and ecosystems are discussed in a
24 number of government documents and in many other publications. They are mentioned here
25 because many of the chemical compounds are partitioned between gas and particle phases in the
26 atmosphere. As particles, they can become airborne, be distributed over a wide area, and affect
27 remote ecosystems. Some of the chemical compounds are of concern because they may reach
28 toxic levels in food chains of both animals and humans; whereas others tend to decrease or
29 maintain the same toxicity as they move through the food chain. Some examples of movement
30 through food chains are provided below.
April 2002 4-74 DRAFT-DO NOT QUOTE OR CITE
-------
1 Many chemical compounds from a variety of anthropogenic sources are released into the
2 ambient air (see Section 4.2.1). In the atmosphere, the emitted compounds initially go through a
3 mixing process, and the airborne particles then are distributed over a wide area and ultimately
4 deposited on ecosystem components. Atmospheric deposition of polychlorinated dibenzo-p-
5 dioxins and dibenzofurans (PCDD/Fs), as an example, can be divided into three different forms:
6 (1) dry gaseous, (2) dry particle-bound, and (3) wet deposition. Dry particle-bound deposition
7 occurs when the PM containing the pollutant is deposited on the plant surface; whereas wet
8 deposition ranges from hail through rain to fog and dew fall (McLachlan, 1996b).
9 Human exposure to PCDD/Fs has been demonstrated to be caused almost exclusively by
10 the ingestion of animal fat from fish, meat, and dairy products. Almost half of human exposure
11 to PCDD/Fs is caused by consumption of beef and dairy products (McLachlan, 1996b). Cattle
12 obtain most of their PCCD/Fs though grass. Therefore, the grass-cattle-milk/beef pathway is
13 critical for human exposure. It has been shown that root uptake/translocation is an insignificant
14 pathway of PCDD/Fs to aerial plant parts. Wet and dry particle deposition are the most
15 important for the accumulation of the higher chlorinated cogeners in vegetation. The persistence
16 of PCDD/Fs in plants has not been investigated extensively; however, biodegradation probably
17 does not occur in that these compounds are found primarily in the lipophilic cuticle and are very
18 resistant to microbial degradation (McLachlan, 1996b). Feed contaminated with soil containing
19 the pollutant can be another source of exposure of beef and dairy cattle, as well as chickens. The
20 PCDD/Fs are near a steady state in milk cows and laying hens; however, animals raised for meat
21 production (such as beef cattle and pigs) may accumulate them. The beef cattle and pigs cannot
22 excrete the contaminants in a lipid-rich matrix such as milk or eggs. Thus, all of the PCDD/Fs,
23 ingested are stored in the body. In agricultural food chains, there is a biodilution of PCDD/Fs,
24 with the fugacity decreasing by up to three orders of magnitude between the air and cows milk
25 (McLachlan, 1996b). Fiirst et al. (1993), based on surveys to determine the factors that influence
26 the presence of PCDD/PCDF in cows milk, earlier concluded that regardless of which pathway,
27 soil - grass - cow or air - grass - cow, it was the congener of the chemical that was most
28 important.
29 Chlorinated persistent organic pollutants (POPS), such as PCBs, PCDFs, and PCDDs, can
30 be transported as particles through the atmosphere from industrial and agricultural sources; be
31 brought down via wet and dry deposition in remote regions, such as the Arctic; and have been
April 2002 4-75 DRAFT-DO NOT QUOTE OR CITE
-------
1 detected in all levels of the Arctic food chain (Oehme et al., 1995). High concentrations of PCB
2 (1 to 10 ppm) were found in seals, but the concentrations increased to 10 to 100 ppm in polar
3 bears. The polar bear is the top predator in the Arctic and feeds preferentially on ringed seals
4 and, to a lesser extent, on other seal species. Bioconcentration factors of organochlorines in the
5 Arctic food web, reaching 107 for fish and seals, are biomagnified in polar bears (Oehme et al.,
6 1995). Polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans have also been
7 found in seals (Oehme et al., 1995). Milk taken from anaesthetized polar bears was also found to
8 contain PCDD/PCDF. Very little is known regarding the intake of milk by polar bear cubs.
9 However, estimates of the intake of milk containing detectable levels of PCDD/PCDF and PCB
10 and the additional consumption of seal blubber confirm that these pollutants are passed on to the
11 next generation (Oehme et al., 1995).
12 Section 112 of the CAA, provides the legislative basis for U.S. hazardous air pollutant
13 (HAP) programs. In response to mounting evidence that air pollution contributes to water
14 pollution, Congress included Section 112m (Atmospheric Deposition to Great Lakes and Coastal
15 Waters) in the 1990 CAA Amendments that direct the EPA to establish a research program on
16 atmospheric deposition of HAPS to the "Great Waters".
17 Actions taken by EPA and others to evaluate and control sources of Great Waters pollutants
18 of concern appear to have positively affected trends in pollutant concentrations measured in air,
19 sediment, and biota. Details concerning these effects may be found in "Deposition of Air
20 Pollutants to the Great Waters", Third Report to Congress (U. S. Environmental Protection
21 Agency, 2000a). The Third Report (EPA-453/R-00-005, June 2000), like the First and Second
22 Reports to Congress, focuses on 15 pollutants of concern, including pesticides, metal
23 compounds, chlorinated organic compounds, and nitrogen compounds. The new scientific
24 information in the Third Report supports and builds on three broad conclusions presented in the
25 previous two EPA Reports to Congress and discussed below.
26 (1) Atmospheric deposition from human activities can be a significant contributor of toxic
27 chemicals and nitrogen compounds to the Great Waters. The relative importance of
28 atmospheric loading for a particular chemical in a water body depends on many factors (e.g.,
29 characteristics of the water body, properties of the chemical, and the kind and amount of
30 atmospheric deposition versus or water discharges).
April 2002 4-76 DRAFT-DO NOT QUOTE OR CITE
-------
1 (2) A plausible link exists between emissions of toxic pollutants of concern into the air above
2 the Great Waters; the deposition of these pollutants (and their transformation products); and
3 the concentrations of these pollutants found in the water, sediments, and biota, especially
4 fish and shellfish. For mercury, fate and transport modeling and exposure assessments
5 predict that the anthropogenic contribution to the total amount of methylmercury in fish is, in
6 part, the result of anthropogenic mercury releases from industrial and combustion sources
7 increasing mercury body burdens (i.e., concentrations) in fish. Also, the consumption offish
8 is the dominant pathway of exposure to methylmercury for fish-consuming humans and
9 wildlife. However, what is known about each stage of this process varies with each pollutant
10 (for instance, the chemical species of the emissions and its transformation in the
11 atmosphere).
12 (3) Airborne emissions from local as well as distant sources, from both within and outside the
13 United States, contribute pollutant loadings to waters through atmospheric deposition.
14 Determining the relative roles of particular sources—local, regional, national, and possibly
15 global, as well as anthropogenic, natural, and reemission of pollutants—contributing to
16 specific water bodies is complex, requiring careful monitoring, atmospheric modeling, and
17 other analytical techniques.
18
19 Ecosystem Response to Indirect Effects of Particulate Matter
20 The presence of PM in the atmosphere directly affects vegetation following physical
21 contact with foliar surfaces (as discussed above), but in many cases the more significant effects
22 are indirect. These effects may be mediated by suspended PM (i.e., through effects on radiation
23 and climate) and by particles that pass through vegetative canopies to reach the soil. Effects
24 mediated in the atmosphere are considered briefly below and in greater detail later, under
25 Section 4.5. Indirect plant responses are chiefly soil-mediated and depend primarily on the
26 chemical composition of the individual elements deposited in PM. The individual elements must
27 be bioavailable to have an effect. The soil environment, composed of mineral and organic
28 matter, water, air, and a vast array of bacteria, fungi, algae, actinomycetes, protozoa, nematodes,
29 and arthropods, is one of the most dynamic sites of biological interactions in nature (Wall and
30 Moore, 1999; Alexander, 1977). The quantity of organisms in soils varies by locality. Bacteria
31 and fungi are usually most abundant in the rhizosphere, the soil around plant roots that all
April 2002 4-77 DRAFT-DO NOT QUOTE OR CITE
-------
1 mineral nutrients must pass through. Bacteria and fungi benefit from the nutrients in the root
2 exudates (chiefly sugars) in the soil and, in turn, they play an essential role by making mineral
3 nutrients available for plant uptake (Wall and Moore, 1999; Rovira and Davey, 1974). Their
4 activities create chemical and biological changes in the rhizosphere by decomposing organic
5 matter and making inorganic minerals available for plant uptake. Bacteria are essential in the
6 nitrogen and sulfur cycles and make these elements available for plant uptake and growth (see
7 Section 4.3.3). Fungi are directly essential to plant growth. Attracted to the roots by the
8 exudates, they develop mycorrhizae, a mutualistic, symbiotic relationship, that is integral in the
9 uptake of the mineral nutrients (Allen, 1991). The impact in ecosystems of PM, particularly
10 nitrates, sulfates, and metals, is determined by their effects on the growth of the bacteria involved
11 in nutrient cycling and the fungi involved in plant nutrient uptake.
12
13 Atmospheric Turbidity: Effects on Vegetative Processes. Photosynthetic processes
14 underlie the contribution of vegetative surfaces to nutrient and energy cycling. Photosynthesis
15 and the heat-driven process of water cycling depend on net receipts and characteristics of the
16 radiation environment. These characteristics may be altered substantially when the atmosphere
17 becomes turbid because of particulate loading. Which wavelengths are of interest depends on the
18 vegetation process under consideration. Canopy temperature and water relations are particularly
19 sensitive to long-wave, infrared radiation; whereas primary photosynthetic charge separations
20 depend on short-wave radiation in the visible and photosynthetically active range (0.4 to 0.7 //m).
21 Effects of anthropogenic aerosols on the radiation environment at the Earth's surface are
22 difficult to assess. The residence time of suspended particles varies with size and environmental
23 conditions (seconds to months or years), and concentrations are spatially and temporally variable.
24 In particularly polluted urban and near-urban areas, unambiguous particulate effects on radiation
25 and local climate may be observed. Visibility was degraded by 50% in a large plume originating
26 in the St. Louis urban area during the midweek, midday period (Pueschel, 1993). In contrast,
27 visibility was reduced by only 20% on weekends when traffic and industrial emissions were
28 reduced. The area affected by the plume from the St. Louis urban area includes highly
29 productive agricultural land.
30 Empirical relationships between the mass of specific components of the aerosol and
31 radiation scattering have been developed (e.g., Pueschel, 1993), from which regional visibility
April 2002 4-78 DRAFT-DO NOT QUOTE OR CITE
-------
1 (or radiation attenuation) isopleths can be constructed if appropriate mass data are available.
2 These estimates support trends observed by direct measurement of turbidity (e.g., Flowers et al.,
3 1969; U.S. Environmental Protection Agency, 1982).
4 Sulfates, nitrates, and elemental carbon dominate effects on visibility, in part, because they
5 frequently dominate the mass profiles and, in part, because they exhibit particularly large
6 absorption coefficients (see Section 4.3). Absorption by particles containing carbon may range
7 from 5 to 10% in rural areas to up to 50% in urban areas (U.S. Environmental Protection Agency,
8 1982). In west-coast cities with contrasting particulate sources and loadings, the common
9 component that related PM to visibility degradation was sulfate between 0.65 and 3.6 //m
10 (Barone et al., 1978). For example, in Los Angeles, sulfate and nitrate had similar effects on
11 visibility (White, 1976), despite the dominance of nitrate from transportation sources in the
12 aerosol, although this is changing with controls on point sources of sulfate (Farber et al., 1994).
13 A long-term global trend of increasing atmospheric optical depth has not been documented
14 (Bolle et al., 1986; Pueschel, 1993) although seasonal and regional effects are substantial. The
15 classic study by Flowers et al. (1969) demonstrated large regional distinctions in turbidity across
16 the United States. Typically, the western deserts, plains, and Rocky Mountains exhibited low
17 mean annual turbidity; whereas the more humid and densely vegetated eastern half of the country
18 exhibited much greater turbidities. In the mid-1970s, visible range in the mountainous southwest
19 exceeded 110 km and radiation attenuation was ca. 2.6%; whereas, in the East, visible range was
20 below 24 km and radiation attenuation was ca. 10%. Visibility in the eastern United States has
21 decreased generally since the 1940s (Flowers et al., 1969; Trijonis and Shapland, 1979; U.S.
22 Environmental Protection Agency, 1982). Correlative trends in visibility degradation and
23 emissions of sulfur oxides suggest that parti culate sulfate may account for much of the turbidity.
24 These trends are typical of urban industrial areas around the world. Turbidity has increased
25 above Mexico City (Binenko and Harshvardhan, 1993) since the 1911 to 1928 period. During
26 this early period, a single annual peak of turbidity coincided with the end of the dry period, and
27 natural sources dominated. By 1957 to 1962, the number of annual peaks had increased as
28 anthropogenic sources came to dominate. During this period, atmospheric transmission of direct-
29 beam solar radiation decreased by about 10% (Binenko and Harshvardhan, 1993). Visibility in
30 the Los Angeles basin has improved very slightly in the past decades (Farber et al., 1994) as
31 sulfate emissions have been controlled by regulation. The composition of the aerosol has
April 2002 4-79 DRAFT-DO NOT QUOTE OR CITE
-------
1 changed, particularly in inland areas, as the former dominance of sulfate shifts to a
2 preponderance of secondary organics.
3 Particles interact with solar radiation through scattering and absorption. Absorption of
4 short-wavelength solar radiation reduces the amount of radiation reaching the Earth's surface and
5 leads to atmospheric heating. If the absorbing particles re-radiate in the infrared range, then
6 some of this energy is lost as long-wave re-radiation to space. This loss mechanism is minimized
7 because most of the anthropogenic aerosol in the troposphere resides in the planetary boundary
8 layer (Bolle et al., 1986), even within the lower 500 m (Binenko and Harshvardhan, 1993) where
9 the temperature is similar to that of the surface. Some of this energy is captured at the surface as
10 down-welling infrared radiation.
11 These wavelengths directly affect canopy temperatures and influence transpirational water
12 use by vegetation. The presence of absorbing aerosols reduces the ratio of photosynthetically
13 active radiation to total radiation received at the surface, potentially reducing photosynthetic
14 water use efficiency. The net effect of aerosol absorption on the surface depends on the relative
15 magnitudes of the particulate absorption coefficients in the visible and infrared area and on the
16 albedo of the Earth's surface. In general, absorption is not a dominant parti culate effect.
17 Scattering of radiation dominates the effects of parti culate loading on visibility and
18 turbidity. Non-absorbing, scattering aerosols raise the overall albedo of the atmosphere and
19 reduce the amount of radiation reaching the surface by the amount reflected or backscattered to
20 space. As atmospheric turbidity increases, so does the scattering of light, including forward
21 scattering of photosynthetically active radiation that intercepts the Earth's surface (Hoyt, 1978).
22 The largest effect is described by Mie-scattering theory. Forward scattering reduces the
23 intensity of direct radiation by disrupting the solar beam, thereby increasing the path length and
24 probability of absorption and by increasing the intensity of diffuse (sky) radiation. In a clear
25 atmosphere, diffuse radiation may be on the order of 10% of total solar radiation (Choudhury,
26 1987). However, in highly turbid, humid conditions, this fraction may increase, even up to 100%
27 of solar radiation in extreme cases. The direct-to-diffuse-radiation ratio is highest at solar noon
28 and lowest near dawn or dusk when the path length through the atmosphere is longest.
29 Particle scattering is wavelength dependent, causing objects to appear blue- or red- tinged
30 depending on viewing and illumination angles and on the light quality, the alteration of which is
31 a minor contributor to photosynthetic light-use efficiency. The wavelength dependence of
April 2002 4-80 DRAFT-DO NOT QUOTE OR CITE
-------
1 scattering decreases rapidly from extreme sensitivity for very fine particles to little dependence at
2 10 (j,m. Equations relating scattering at a reference wavelength to scattering at wavelengths of
3 interest are rigorously applicable only to spherical particles but may be extended to nonspherical
4 particles of equal volume (Janzen, 1980).
5 World Meteorological Organization data summarized in U.S. Environmental Protection
6 Agency (1982) indicated that turbidity in the eastern United States commonly resulted in
7 radiation losses of ca. 3.5% because of backscattered radiation and ca. 3.5% because of
8 absorption, with a resulting total reduction of incident radiation to ca. 93% of total solar
9 radiation. However, 28% of the radiation reaching the surface was converted from direct
10 radiation to diffuse, or sky, radiation. Under more polluted conditions, losses were ca. 9%
11 backscattered and 9% absorbed, reducing total radiation to 82% of total solar radiation and
12 converting 72% from direct beam to diffuse radiation. Photosynthetically active radiation (0.4 to
13 0.7 //m) typically is enriched in diffuse radiation relative to total or direct beam radiation.
14
15 Altered Radiative Flux: Effects on Vegetative Processes. Canopy photosynthesis is
16 typically a nearly linear function of incident radiation, overcoming saturation exhibited by
17 individual leaves by distributing the light throughout the multilayer canopy. Light penetration
18 into canopies limits photosynthetic productivity (Rosenberg et al., 1983). The uppermost leaves
19 of many canopies are at or above light saturation for photosynthetic processes. The simplest
20 radiative transfer functions describing plant canopies relate total down-welling radiation (direct
21 plus diffuse radiation measured above the canopy) to radiation interception at each leaf level
22 through a Beer's Law analogy. The expected exponential decline in radiation through the canopy
23 depends only on total radiation and a bulk canopy extinction coefficient that depends on leaf size,
24 orientation, and distribution, as well as on reflectance and absorption in wavelengths of interest.
25 These simplified models predict radiation distribution adequately for homogeneous canopies.
26 Turbidity affects canopy processes only by attenuating the total radiation impinging on the
27 canopy surface.
28 In more complex, and more realistic, canopy-response models (e.g., Choudhury, 1987),
29 radiation is considered in its direct and diffuse components. Foliar interception by canopy
30 elements is considered for both up- and down-welling radiation (a two-stream approximation).
31 In this case, the effect of atmospheric PM on turbidity affects canopy processes both by radiation
April 2002 4-81 DRAFT-DO NOT QUOTE OR CITE
-------
1 attenuation and by influencing the efficiency of radiation interception throughout the canopy
2 through conversion of direct to diffuse radiation (Hoyt, 1978). Diffuse radiation is more
3 uniformly distributed throughout the canopy and increases canopy photosynthetic productivity by
4 distributing radiation to lower leaves. The treatment of down-welling direct-beam radiation in
5 the two-stream approach remains an elaboration of the simplified Beer's Law analogy with solar
6 angle, leaf area distribution, and orientation individually parameterized (Choudhury, 1987).
7 Diffuse down-welling radiation is a function of diffuse and direct radiation at the top of the
8 canopy and penetration within the canopy according to cumulative leaf area density and foliage
9 orientation. Up-welling (diffuse) radiation results from scattering and reflectance of both direct
10 and diffuse down-welling radiation within the canopy and by the soil.
11 The altered distribution between diffuse and direct radiation affects photosynthesis in
12 upper, exposed leaves as a function of leaf angle and in total canopy photosynthesis as a function
13 of penetration of radiation within the canopy. This depends on canopy structure, leaf optical
14 properties, and leaf area density, as well as on solar angle and atmospheric turbidity. Absorption
15 of radiation by particles heats the upper atmosphere and results in reduced vertical temperature
16 gradients. This could reduce the intensity of atmospheric turbulent mixing. The magnitude of
17 such potential effects on turbulent transport within canopies remains unknown although damping
18 of eddy transport could inhibit canopy gas exchange. Suppressed tropospheric mixing also could
19 intensify local temperature inversions and increase the severity of pollution episodes (Pueschel,
20 1993) with direct inhibitory effects on photosynthetic processes.
21 The most significant effect of aerosols on vegetation is probably through their role as cloud
22 condensation nuclei because clouds have a substantial effect on radiation receipts at the surface.
23 An important characteristic of fine particles is their ability to affect the flux of solar radiation
24 passing through the atmosphere directly, by scattering and absorbing solar radiation, and
25 indirectly, by acting as cloud condensation nuclei which in turn influence the optical properties
26 of clouds (Chameides et al., 1999). Regional haze has been estimated to diminish surface solar
27 visible radiation by approximately 8%. Crop yields have been reported as being sensitive to the
28 amount of sunlight received. The potentially significant effect of regional haze on the yield of
29 crops because of reduction in solar radiation has been examined by Chameides et al. (1999).
30 Using a case study approach, Chameides et al. (1999), studied the effects of regional haze on
31 crop production in China where regional haze is especially severe. A rudimentary assessment of
April 2002 4-82 DRAFT-DO NOT QUOTE OR CITE
-------
1 the direct effect of atmospheric aerosols on agriculture suggests that optimal crop yields of
2 approximately 70% of the crops are being depressed by at least 3 to 5% by regional scale air
3 pollution and its associated haze (Chameides et al., 1999).
4
5 Effects of Solar Ultraviolet Radiation. The transmission of solar UV-B radiation through
6 the earth's atmosphere is controlled by ozone, clouds and particles. The depletion of
7 stratospheric ozone, caused by the release of chlorofluorcarbons (CFCs) and other substances
8 such as halides, has resulted in heightened concern about potentially deleterious increases in the
9 amount of solar UV-B (SUVB) radiation reaching the Earth's surface. One salient consideration
10 is that, although CFC levels in the stratosphere have reached peak levels and now are beginning
11 to fall as a result of the signing of the Montreal Protocol, the problem will likely continue well
12 into the future because of the length of time it takes ozone-depleting molecules to reach the
13 stratosphere (Greenberg, 1997).
14 The vulnerability of terrestrial plants to UV-B results from their requirement for sunlight
15 for photosynthesis. Each 1% decline in stratospheric ozone has been predicted to decrease crop
16 yield by 1% (Greenberg et al., 1997). In addition to inhibiting photosynthesis, UV-B radiation
17 triggers numerous responses in plants, e.g.: membrane, protein, and DNA damage; delayed
18 maturation; diminished growth; activation of chemical stress; flavonoid synthesis; and leaf
19 thickening (Table 4-15). It is not known which of the injury and damage effects are most
20 detrimental to plant growth (Table 4-15). Effects of increased UV-B on plant growth are likely
21 to be incremental. Because plants evolved under the selective pressure of ambient UV-B
22 radiation in sunlight, they have developed adaptive mechanisms (Greenberg et al., 1997).
23 Although inhibition of photosynthesis is a detrimental growth effect, flavonoid synthesis
24 represents acclimation. Plants growing under full light have been shown to be protected against
25 UV-B effects, but not when growing under weak visible light (Bjorn, 1996). A common
26 adaptation is alteration in leaf transmission properties, which results in attenuation of UV-B in
27 the epidermis before it can reach the leaf interior.
28 Plant species vary enormously in their response to UV-B exposures, and large differences
29 in response occur among different genotypes within a species. In general, dicotyledonous plants
30 are more sensitive than monocotyledons from similar environments. In addition, plant responses
31 may differ depending on stage of development. Therefore, extrapolation of experimental
April 2002 4-83 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-15. TYPES OF PLANT RESPONSES TO ULTRAVIOLET-B RADIATION3
Acclimation and Morphological Responses
Damage and Injury Responses
Altered biomass distribution
Altered leaf cell division
Cotyledon curling
Increased DNA repair
Increased flavonoid biosynthesis
Increased leaf thickness
Increased leaf number
Increased number of tillars
Leaf wrinkling
Reduced leaf area
Reduced hypocotyl growth
Reduced shoot height
Reduced stomatal density
Altered gene expression
Degradation of auxin
Degradation of chlorophyll and carotenoids
Degradation of proteins
Diminished biomass
Epidermal collapse
Inhibition of growth
Inhibition of photosynthesis
Increased stomatal conductance
Lower seed yield
Oxidation of DNA
Peroxidation of lipids
Prymidine dimer formation
"Entries in alphabetical order.
1 responses from seedlings to mature plants must be made with caution (Bjorn, 1996). The above
2 facts are especially important when considering the effects of UV-B on agricultural plants.
3 For example, among soybeans and rice, there are varieties for which growth and crop yield are
4 severely decreased by increased UV-B radiation and other varieties that are not affected or may
5 even be stimulated. On the other hand, the growth of the same sensitive soybeans when grown
6 under water stress was not inhibited. Many crop plants grown in temperate regions originated in
7 more tropical areas, hence, a gene pool for more resistant varieties is likely to exist (Bjorn, 1996).
8 Crop plants, unlike forest trees and vegetation in natural ecosystems, are only exposed for one
9 generation; and, thus, it may be possible to readily change the genotype if a variety proves to be
10 sensitive.
April 2002
4-84
DRAFT-DO NOT QUOTE OR CITE
-------
1 Trees, forests, and perennial evergreen plants are long-lived when compared to agricultural
2 systems, making it possible for UV-B exposure impacts to accumulate with time. Saplings and
3 young and small trees react differently when compared to mature trees; also, on evergreen trees,
4 needles of different ages respond differently (Bjorn, 1996). Breeding and testing trees is a slow
5 process; and, for this reason, much care needs to be taken when planting large areas with trees of
6 a single species and one provenance (e.g., Stika Spruce [Picea sitchensis] in Britain). The
7 response of only a few broad-leaved trees have been studied. The most investigated genus has
8 been loblolly pine (Pinus taeda; Bjorn, 1996).
9 A few studies indicate that the photomorphogenesis (changes in leaf thickness under UV-B
10 that results in a transition from shade to sun leaves, Table 4-15) and the variable responses of
11 native plants in ecosystems to UV-B exposures results in changes in interactions between various
12 plants species, changes between plants and other organisms, and between plants and their abiotic
13 environment. These preliminary studies suggest that in natural ecosystems, composed of many
14 different plant species, with complex interactions between plants and between plants and other
15 organisms, effects of UV-B may develop that cannot be determined from experiments on single
16 plant species. The effects of UV-B on natural plant systems, therefore, should be of greater
17 concern than on agricultural crops (Bjorn, 1996).
18
19 Effects of Nitrogen Deposition. Nitrogen has long been recognized as the nutrient most
20 important for plant growth. Nitrogen is of overriding importance in plant metabolism and, to a
21 large extent, governs the utilization of phosphorus, potassium, and other nutrients. Most of the
22 nitrogen in soils is associated with organic matter. Typically, the availability of nitrogen via the
23 nitrogen cycle controls net primary productivity, and possibly, the decomposition rate of plant
24 litter. Photosynthesis is influenced by nitrogen uptake in that ca. 75% of the nitrogen in a plant
25 leaf is used during the process of photosynthesis. The nitrogen-photosynthesis relationship is,
26 therefore, critical to the growth of trees and other plants (Chapin et al., 1987). Plants usually
27 obtain nitrogen directly from the soil through their roots by absorbing NH4 + or NO3", or it is
28 formed by symbiotic organisms in the roots. Plants, however, vary in their ability to absorb
29 ammonium and nitrate (Chapin et al., 1987).
30 Because nitrogen is not readily available and is usually in shortest supply, it is the chief
31 element in agricultural fertilizers. Atmospherically deposited nitrogen also can act as a fertilizer
April 2002 4-85 DRAFT-DO NOT QUOTE OR CITE
-------
1 in soil low in nitrogen. Not all plants, however, are capable of utilizing extra nitrogen. Inputs of
2 nitrogen to natural ecosystems that alleviate deficiencies and increase growth of some plants can
3 alter competitive relationships and alter species composition and diversity (Ellenberg, 1987;
4 Kenk and Fischer, 1988; U.S. Environmental Protection Agency, 1993).
5 The effect of increasing nitrogen inputs (e.g., NOX, nitrates, nitric acid) on the nitrogen
6 cycle in forests, wetlands, and aquatic ecosystems is discussed in detail elsewhere (U.S.
7 Environmental Protection Agency, 1993, 1997'a; Garner, 1994; World Health Organization,
8 1997). The sources and forms of organic nitrogen in the atmosphere are poorly studied, and the
9 concentrations are rarely measured, except in precipitation. Possible sources include particulate-
10 entrained material from soils and vegetation (e.g., pollen, soil dust and spores) and reaction
11 products of nitrogen oxides with organic compounds (e.g., peroxyacetyl nitrate, PAN; Lovett,
12 1992). The most important effects of nitrogen deposition are accumulation of nitrogen
13 compounds resulting in the enhanced availability of nitrate or ammonium, soil-mediated effects
14 of acidification, and increased susceptibility to stress factors (Bobbink et al., 1998). A major
15 concern is "nitrogen saturation," the result of the deposition of large amounts of particulate
16 nitrates. Nitrogen saturation results when additions to soil background nitrogen (nitrogen
17 loading) exceeds the capacity of plants and soil microorganisms to utilize and retain nitrogen
18 (Aber et al., 1989, 1998; Garner, 1994; U.S. Environmental Protection Agency, 1993). Under
19 these circumstances, disruptions of ecosystem functioning may result (Hornung and Langan,
20 1999).
21 The growth of most forests in North America is limited by the nitrogen supply. Severe
22 symptoms of nitrogen saturation, however, have been observed in high-elevation, nonaggrading
23 spruce-fir ecosystems in the Appalachian Mountains, as well as in the eastern hardwood
24 watersheds at Fernow Experimental Forest near Parsons, WV. Mixed conifer forests and
25 chaparral watersheds with high smog exposure in the Los Angeles Air Basin also are nitrogen
26 saturated and exhibit the highest stream water NO3" concentrations for wildlands in North
27 America (Bytnerowicz and Fenn, 1996; Fenn et al., 1998). Forests in southern California, the
28 southwestern Sierra Nevada in Central California, and the Front Range in northern Colorado
29 have all been exposed to highly elevated nitrogen deposition, and nitrogen saturated watersheds
30 have been reported in the above mentioned areas. Annual nitrogen additions through deposition
31 (6-11 kg ha"1 y"Jas through fall) in the southwestern Sierra Nevada are similar to nitrogen storage
April 2002 4-86 DRAFT-DO NOT QUOTE OR CITE
-------
1 (4 kg ha"1 y"1) in vegetation growth increment of western forests suggesting that current nitrogen
2 deposition rates may be near the assimilation capacity of the overstory vegetation. Ongoing
3 urban expansion will increase the potential for nitrogen saturation of forests from urban sources
4 (e.g., Salt Lake City, Seattle, Tucson, Denver, central and southern California) unless there are
5 improved emission controls (Fenn et al., 1998).
6 Not all forest ecosystems react in the same manner to nitrogen deposition. High-elevation
7 alpine watersheds in the Colorado Front Range (Bowman, 2000) and a deciduous forest in
8 Ontario, Canada, also are naturally saturated even though nitrogen deposition has been moderate
9 (~8 kg ha"1 y"1). The nitrogen saturated forests in North America, including estimated inputs and
10 outputs, are shown in Table 4-16 (Fenn et al., 1998). The Harvard Forest hardwood stand in
11 Massachusetts, however, has absorbed >900 kg N/ha without significant NO3" leaching during a
12 nitrogen amendment study of 8 years (Table 4-16; Fenn et al.,1998). Johnson et al. (1991a)
13 reported that measurements showing the leaching of nitrates and aluminum (Al+3) from high
14 elevation forests in the Great Smoky Mountains indicate that these forests have reached
15 saturation.
16 Possible ecosystem responses to nitrate saturation, as postulated by Aber and coworkers
17 (Aber et al., 1989), include (1) a permanent increase in foliar nitrogen and reduced foliar
18 phosphorus and lignin caused by the lower availability of carbon, phosphorus, and water;
19 (2) reduced productivity in conifer stands because of disruptions of physiological function;
20 (3) decreased root biomass and increased nitrification and nitrate leaching; and (4) reduced soil
21 fertility, resulting from increased cation leaching, increased nitrate and aluminum concentrations
22 in streams, and decreased water quality. Saturation implies that some resource other than
23 nitrogen is limiting biotic function.
24 Water and phosphorus for plants and carbon for microorganisms are the resources most
25 likely to be the secondary limiting factors. The appearance of nitrogen in soil solution is an early
26 symptom of excess nitrogen. In the final stage, disruption of forest structure becomes visible
27 (Garner, 1994).
28 Changes in nitrogen supply can have a considerable effect on an ecosystem's nutrient
29 balance (Waring, 1987). Large chronic additions of nitrogen influence normal nutrient cycling
30 and alter many plant and soil processes involved in nitrogen cycling (Aber et al., 1989).
31 Among the processes affected are (1) plant uptake and allocation, (2) litter production,
April 2002 4-87 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-16. NITROGEN-SATURATED FORESTS IN NORTH AMERICA,
INCLUDING ESTIMATED N INPUTS AND OUTPUTS
to
o
o
to
Location
Forest Type
Elevation
(m)
N Input
(kg ha"1 year"1)
N Output
(kg ha"1 year"1)
Reference
OO
OO
O
&
O
o
2
o
H
O
c
o
H
W
O
?d
O
Adirondack Mts. northeastern New York
Catskill Mts., southeastern New York
Turkey Lakes Watershed, Ontario, Canada
Whitetop Mt., southwestern Virginia
Femow, West Virginia
Great Smoky Mts. National Park,
Tennessee
Great Smoky Mts. National Park,
Becking Site, North Carolina
Great Smokey Mts. National Park,
Tower Site, North Carolina
Front Range, Colorado
San Dimas, San Gabriel Mts.
southern California
Camp Paivika, San Bernadino Mts.,
southern California
Northern hardwoods or hardwood/ 396-661
conifer mix
Mainly hardwood; some eastern hemlock 335-675
Sugar maple and yellow birch 350-400
Red spruce 1650
Mixed hardwood 735-870
American beech 1600
Red spruce 1800
Red spruce 1740
Alpine tundra, subalpine conifer 3000-4000
Chapparral and grasslands 580-1080
Mixed conifer 1600
9.3"
Stage 1 N lossb Driscoll and Van Dreason (1993)
10.2" Stage 1 and 2 N lossb Stoddard (1994)
7.0-7.7 (as throughfall) 17.9-23.6 Foster et al. (1989); Johnson and Lindberg (1992a)
32C 47C Joslin and Wolfe (1992); Joslinetal. (1992)
15-20 6.1 Gilliam et al. (1996); Peterjohn et al. (1996)
3. ld 2.9 Johnson and Lindberg (1992b)
10.3d
26.6
7.5-8.0
23.3e
30
19.2 Johnson etal. (1991a)
20.3 Johnson etal. (1991a)
7.5 Williams etal. (1996)
0.04-19.4 Riggan et al. (1985)
7-26f
Fenn etal. (1996)
Klamath Mts, northern California
Thompson Forest, Cascade Mts.,
Washington
Western coniferous
Red alder
NA
220
Mainly geologic8
4.7 plus > 100asN2
fixation
NA8
38.9
Dahlgren (1994)
Johnson and Lindberg (1992b)
"Estimated total N deposition from wet deposition data is from Driscoll et al. (1991) for the Adirondacks, and from Stoddard and Murdoch (1991) for the Catskills. Total deposition was estimated based
on the wet deposition/total N deposition ratio (0.56) at Huntington Forest in the Adirondacks (Johnson and Lindberg, 1992b). Nitrogen deposition can be higher in some areas, especially at high-elevation
sites such as Whiteface Mountain (15.9 kg ha"1 year"1; Johnson and Lindberg, 1992b).
""Stage 1 and 2 of N loss according to the watershed conceptual model of Stoddard (1994). Nitrogen discharge (kg ha"1 year"1) data are not available; only stream water NO3" concentration trend data were
collected.
"Values appear high compared to other sites, especially N leaching losses. Joslin and Wolfe (1992) concede that "there is considerable uncertainty associated with the estimates of atmospheric
deposition and leaching fluxes." However, elevated NO3" concentrations in soil solution, and lack of a growth response to N fertilization (Joslin and Wolfe, 1994) support the hypothesis that the forest
at Whitetop Mountain is N saturated.
dEstimated total N deposition from throughfall data. Total deposition was estimated based on the throughfall/total N deposition ration (0.56) from the nearby Smokies Tower site (Johnson and Lindberg,
1992b).
'Annual throughfall deposition to the chaparral ecosystem.
'Nitrogen output is from unpublished streamwater data (M.E. Fenn and M.A. Poth, 1999). The low value represents a year of average precipitation, and the high value is for 1995, when precipitation was
nearly double the long-term average. Nitrogen output includes N export in streamwater and to groundwater.
8Annual input and output data are not known, although N deposition in this forest is probably typical for much of the rural western United States (2-3 kg N ha"1 year"1 (Young et al., 1988). Excess N is
from weathering of ammonium in mica schist bedrock. The ammonium was rapidly nitrified, leading to high NO3" concentrations in soil solution (Dahlgren, 1994).
-------
1
2
3
4
5
(3) immobilization (includes ammonification [the release of ammonia] and nitrificatrion
[conversion of ammonia to nitrate during decay of litter and soil organic matter]), and (4) nitrate
leaching and trace gas emissions (Figure 4-8; Aber et al., 1989).
Deposition
/
Plant
Utilization
S*
Photosynthesis
X
Animal
Proteins
Process altered by
nitrogen saturation
Figure 4-8. Nitrogen cycle (dotted lines indicate processes altered by nitrogen saturation).
Source: Garner (1994).
1 Subsequent studies have shown that, although there was an increase in nitrogen
2 mineralization initially (i.e., the conversion of soil organic matter to nitrogen in available form
3 [see item 3 above]), nitrogen mineralization rates were reduced under nitrogen-enriched
April 2002
4-89
DRAFT-DO NOT QUOTE OR CITE
-------
1 conditions. Also, studies suggest that soil microbial communities change from predominantly
2 fungal (mycorrhizal) communities to those dominated by bacteria during saturation (Aber et al.,
3 1998).
4 Because the competitive equilibrium of plants in any community is finely balanced, the
5 alteration of one of a number of environmental parameters, (e.g., continued nitrogen additions),
6 can change the vegetation structure of an ecosystem (Bobbink, 1998; Skeffmgton and Wilson,
7 1988). Increases in soil nitrogen play a selective role. When nitrogen becomes more readily
8 available, plants adapted to living in an environment of low nitrogen availability will be replaced
9 by plants capable of using increased nitrogen because they have a competitive advantage.
10 Plant succession patterns and biodiversity are affected significantly by chronic nitrogen
11 additions in some North American ecosystems (Figure 4-9). The location of nitrogen saturated
12 ecosystems in North America, and the steps leading to nitrogen saturation, are indicated on the
13 map in Figure 4-9. Conceptual models of regional nitrogen saturation are located in New
14 England, the Colorado alpine ecosystems and in California forests. Fenn et al. (1998) report that
15 long-term nitrogen fertilization studies in both New England and Europe, as well, suggest that
16 some forests receiving chronic inputs of nitrogen may decline in productivity and experience
17 greater mortality. Long-term fertilization experiments at Mount Ascutney, Vermont, suggest that
18 declining coniferous forest stands with slow nitrogen cycling may be replaced by deciduous
19 fast-growing forests that cycle nitrogen rapidly (Fenn et al., 1998).
20 Competition among species can result in changes in community composition; therefore, it
21 is one of the most notable responses to environmental change (Bowman, 2000). Nitrogen
22 saturation, the result of increased deposition in the alpine tundra of Niwot Ridge in the Front
23 Range of the Southern Rockies in Colorado has changed nitrogen cycling and provided the
24 potential for replacement in plant species by more competitive, faster growing species (Bowman
25 and Steltzer, 1998; Bowman, 2000; Baron et al., 2000). Plants growing in an alpine tundra, as is
26 true of other plants growing in low resource environments (e.g., infertile soil, shaded understory,
27 deserts), have been observed to have certain similar characteristics: a slow grow rate, low
28 photosynthetic rate, a low capacity for nutrient uptake and low soil microbial activity (Bowman
29 and Steltzer, 1998; Bowman, 2000). An important feature of such plants is that they continue to
30 grow slowly and tend to respond even less when provided with an optimal supply and balance of
31 resources (Pearcy et al., 1987; Chapin, 1991). Plants adapted to cold, moist environments grow
April 2002 4-90 DRAFT-DO NOT QUOTE OR CITE
-------
N-Saturated Ecosystems
in North America
Review of Ecosystem Effects
and Responses to Excess N
1. Nitrogen Inputs:
>• Atmospheric deposition, N2 fixation, fertilization
Nitrogen Retention:
> In plant biomass and soil organic matter
> The role of soil microbes and woody residues
> Abiotic retention
Nitrogen Outputs:
>-Hydrologic transport, gaseous emissions from soil
> Removal in harvest, fire emissions, and soil erosion
2. Characteristics Predisposing Forests
to N Saturation:
>• Stand vigor and succession, forest type
> Previous land use-stand history
> Soil N accumulation
> Topography and climate
> Nitrogen deposition
3. Ecosystem Responses to Excess Nitrogen:
> Nitrate leaching and export
> Eutrophicationof estuaries
>• Toxicity of surface waters
>• Foliar nutrient responses
>• Nitrogen mineralization and nitrification
> Effects on soil organic matter
>• Soil acidification, cation depletion, Al toxicity
> Foliar nutrient responses
> Greenhouse gas fluxes
4. Regional N Saturation Conceptual Models:
> New England forests
> California forests
> Colorado alpine ecosystems
Figure 4-9. Diagrammatic overview of excess nitrogen (N) in North America.
Source: Fennetal. (1998).
April 2002
4-91
DRAFT-DO NOT QUOTE OR CITE
-------
1 more leaves than roots as the relative availability of nitrogen increases; however, other nutrients
2 may soon become limiting. These patterns of vegetative development affect their capacity to
3 respond to variation in available resources and to environmental stresses such as frost, high
4 winds, and drought. Preformation of buds 3-4 years in advance of emergence, reduced cell
5 numbers, and high biomass allocation to belowground organs also limits the ability of many
6 alpine plants to respond to variations in their environment (Bowman, 2000). However,
7 significant interspecific genetic variation influences the capacity of the alpine species to respond
8 to changes in resource availability. The capacity of subalpine and boreal species in particular,
9 and gymnosperms in general, to reduce nitrates in either roots or leaves appears to be limited. In
10 addition, the ability of trees to use nitrogen varies with the age of the tree and the density of the
11 stand (Waring, 1987).
12 In experimental studies of nitrogen deposition conducted by Wedin and Tilman (1996) over
13 a 12-year period on Minnesota grasslands, plots dominated by native warm-season grasses
14 shifted to low-diversity mixtures dominated by cool-season grasses at all but the lowest rates of
15 nitrogen addition. Grasslands with high nitrogen retention and carbon storage rates were the
16 most vulnerable to loss of species and major shifts in nitrogen cycling. The shift to low-diversity
17 mixtures was associated with the decrease in biomass carbon to nitrogen (C:N) ratios, increased
18 nitrogen mineralization, increased soil nitrate, high nitrogen losses, and low carbon storage
19 (Wedin and Tilman, 1996). Naeem et al. (1994) experimentally demonstrated (under controlled
20 environmental conditions) that the loss of biodiversity, genetic resources, productivity, ecosystem
21 buffering against ecological perturbation, and loss of aesthetic and commercially valuable
22 resources also may alter or impair ecosystems services.
23 The long-term effects of increased nitrogen deposition have been studied in several western
24 and central European plant communities: lowland heaths, species-rich grasslands, mesotrophic
25 fens, ombrotrophic bogs, upland moors, forest-floor vegetation, and freshwater lakes (Bobbink,
26 1998). Large changes in species composition have been observed in regions with high nitrogen
27 loadings or infield experiments after years of nitrogen addition (Bobbink et al., 1998). The
28 increased input of nitrogen gradually increased availability of nitrogen in the soil, and its
29 retention because of low rates of leaching and denitrification resulted in faster litter
30 decomposition and rate of mineralization. Faster growth and greater height of nitrophilic species
31 enables these plants to shade out the slower growing species, particularly those in oligotrophic or
April 2002 4-92 DRAFT-DO NOT QUOTE OR CITE
-------
1 mesotrophic conditions (Bobbink, 1998; Bobbink et al., 1998). Excess nitrogen inputs to
2 unmanaged heathlands in the Netherlands has resulted in nitrophilous grass species replacing
3 slower growing heath species (Roelofs et al., 1987; Garner, 1994). Van Breemen and Van Dijk
4 (1988) noted that over the past several decades the composition of plants in the forest herb layers
5 has been shifting toward species commonly found on nitrogen-rich areas. It also was observed
6 that the fruiting bodies of mycorrhizal fungi had decreased in number.
7 Other studies in Europe point out the effects of excessive nitrogen deposition on mixed-oak
8 forest vegetation along a deposition gradient largely controlled by soil acidity, nitrogen supply,
9 canopy composition, and location of sample plots (Brunet et al., 1998; Falkengren-Grerup,
10 1998). Results of the study, using multivariate methods, suggest that nitrogen deposition has
11 affected the field-layer vegetation directly by increased nitrogen availability and, indirectly, by
12 accelerating soil acidity. Time series studies indicate that 20 of the 30 field-layer species
13 (nonwoody plants) that were associated most closely with high nitrogen deposition increased in
14 frequency in areas with high nitrogen deposition during the past decades. Included in the field-
15 layer species were many generally considered nitrophilous; however, there were several acid
16 tolerant species (Brunet et al, 1998). Falkengren-Grerup (1998), in an experimental study
17 involving 15 herbs and 13 grasses, observed that species with a high nitrogen demand and a
18 lesser demand for other nutrients were particularly competitive in areas with acidic soils and high
19 nitrogen deposition. The grasses grew better than herbs with the addition of nitrogen. It was
20 concluded that, at the highest nitrogen deposition, growth was limited for most species by the
21 supply of other nutrients; and, at the intermediate nitrogen concentration, the grasses were more
22 efficient than the herbs in utilizing nitrogen. Nihlgard (1985) suggested that excessive nitrogen
23 deposition may contribute to forest decline in other specific regions of Europe. Also, Schulze
24 (1989), Heinsdorf (1993), and Lamersdorf and Meyer (1993) attribute magnesium deficiencies in
25 German forests, in part, to excessive nitrogen deposition.
26 The carbon to nitrogen (C:N) ratio of the forest floor can also be changed by nitrogen
27 deposition over time. This change appears to occur when the ecosystem becomes nitrogen
28 saturated (Gundersen et al., 1998a). Long-term changes in C:N status have been documented in
29 Central Europe and indicate that nitrogen deposition has changed the forest floor. In Europe, low
30 C:N ratios coincide with high deposition regions (Gundersen et al., 1998a). A strong decrease in
31 forest floor root biomass has been observed with increased nitrogen availability. Roots and the
April 2002 4-93 DRAFT-DO NOT QUOTE OR CITE
-------
1 associated mycorrhizae appear to be an important factor in the accumulation of organic matter in
2 the forest floor at nitrogen limited sites. If root growth and mycorrhizal formation are impaired
3 by nitrogen deposition, the stability of the forest floor may be affected by stimulating turnover
4 and decreasing the root litter input to the forest floor and thus decrease the nitrogen that can be
5 stored in the forest floor pool (Gundersen et al., 1998b). Nitrogen-limited forests have a high
6 capacity for deposited nitrogen to be retained by the plants and microorganisms competing for
7 available nitrogen (Gundersen et al., 1998b). Nitrate leaching has been correlated significantly
8 with nitrate status but not with nitrate depositions. Forest floor C:N ratio has been used as a
9 rough indicator of ecosystem nitrogen status in mature coniferous forests and the risk of nitrate
10 leaching; analyses of European databases indicated an empirical relationship between forest floor
11 C:N ratio and nitrate leaching (Gundersen et al., 1998a). Nitrate leaching was observed when the
12 deposition received was more than 10 kg N/ha. All of the data sets supported a threshold at
13 which nitrate leaching seems to increase at a C:N ratio of 25. Therefore, to predict the rate of
14 changes in nitrate leaching, it is necessary to be able to predict the rate of changes in the forest
15 floor C:N ratio. Decreased foliar and soil nitrogen and soil C:N ratios, as well as changes in
16 nitrogen mineralization rates, have been observed when comparing responses to nitrogen
17 deposition in forest stands east and west of the Continental Divide in the Colorado Front Range
18 (Baron et al., 2000; Rueth and Baron, 2002). Understanding the variability in forest ecosystem
19 response to nitrogen input is essential in assessing pollution risks (Gundersen et al., 1998a).
20 The plant root is an important region of nutrient dynamics. The rhizosphere includes the
21 soil that surrounds and is influenced by plant roots (Wall and Moore, 1999). The mutualistic
22 relationship between plant roots, fungi, and microbes is critical for the growth of the organisms
23 involved. The plant provides shelter and carbon; whereas the symbiont provides access to a
24 limiting nutrients such as nitrogen and phosphorus. As indicated above, changes in soil nitrogen
25 influence the mycorrhizal-plant relationship. Mycorrhizal fungal diversity is associated with
26 above-ground plant biodiversity, ecosystem variability, and productivity (Wall and Moore, 1999).
27 Aber et al. (1998) showed a close relationship between mycorrhizal fungi and the conversion of
28 dissolved inorganic nitrogen to soil nitrogen. During nitrogen saturation, soil microbial
29 communities change from being fungal, and probably being dominated by mycorrhizae, to being
30 dominated by bacteria. The loss of mycorrhizal function has been hypothesized as the key
April 2002 4-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 process leading to increased nitrification and nitrate mobility. Increased nitrate mobility leads to
2 increased cation leaching and soil acidification (Aber et al., 1998).
3 The interrelationship of above- and below-ground flora is illustrated by the natural invasion
4 of heathlands by oaks (Quercus robuf). Soils are dynamic entities, the features of which can
5 change like the rest of the ecosystem with age and management. The soil-forming factors under
6 the heath have been vegetation typed during the last 2000 years; whereas the invasion by oaks
7 has been taking place for only a few decades. Clearly changes in the ground floor and soil
8 morphology takes place when trees colonize heath (Nielsen et al., 1999). The distribution of
9 roots also changed under the three different vegetation types. Under both heather and the Sitka
10 spruce plantation, the majority of roots are confined to the uppermost horizons; whereas under
11 oak, the roots are distributed more homogeneously. There was also a change in the C:N ratio
12 when heather was replaced by oaks. Also, the spontaneous succession of the heath by oaks
13 changed the biological nutrient cycle into a deeper vertical cycle when compared to the heath
14 where the cycle is confined to the upper soil horizons. Soils similar to those described in this
15 study (Jutland, Denmark) with mainly an organic buffer system seem to respond quickly to
16 changes in vegetation (Nielsen et al., 1999).
17 The affects of changes in root to shoot relationships in plants were observed in studies of
18 the coastal sage scrub community in southern California which is composed of the drought-
19 deciduous shrubs Artemisia californica, Enceliafarinosa, and Eriogonumfasciculatum. The
20 coastal sage scrub in California has been declining in land area and in shrub density over the past
21 60 years and is being replaced in many areas by Mediterranean annual grasses (Allen et al., 1998;
22 Padgett et al., 1999; Padgett and Allen, 1999). Nitrogen deposition was considered as a possible
23 cause. Up to 45 kg/ha/yr are deposited in the Los Angeles Air Basin (Bytnerowicz and Fenn,
24 1996). Tracts of land set aside as reserves, which in many cases in southern California are
25 surrounded by urbanization, receive large amounts of nitrogenous compounds from polluted air.
26 The coastal sage scrub is of particular interest because some 200 sensitive plant species and
27 several federally listed animal species are found in the area (Allen et al., 1998). Because changes
28 in plant community structure often can be related to increases in the availability of a limiting soil
29 nutrient or other resource, experiments were conducted to determine whether increased nitrogen
30 availability was associated with the significant loss in native shrub cover. Studies indicated that
31 the three native perennial shrubs (Artemisia californica, Eriogonumfasciculatum, and Encelia
April 2002 4-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 farinosa tended to be more nitrophilous than the two exotic annual grasses (Bromus mbens,
2 Avenafatud) and the weedy pod mustard (Brassica geniculatd). These results contrast with most
3 models dealing with the adaptation of perennial species to stressful environments (Padgett and
4 Allen, 1999). If nitrogen were the only variable between the invasive annuals and native shrubs,
5 neither shrubs nor grasses have a particular advantage. However, additional studies indicated
6 that the decline in the coastal sage scrub was not associated with its inability to compete with the
7 grasses, but rather with changes in the arbuscular mycorrhizal community in the soil (Edgerton-
8 Warburton and Allen, 2000). Nitrogen enrichment of the soils induced a shift in the arbuscular
9 mycorrhizal community composition. Larger-spored fungal species (Scutellospora and
10 Gigaspora), due to a failure to sporulate, decreased in number with a concomitant proliferation
11 of small-spored species of Glomus aggregatum, G. leptotichum, and G. geospomm, indicating a
12 strong selective pressure for the smaller spores species of fungi (Edgerton-Warburton and Allen,
13 2000). These results demonstrate that nitrogen enrichment of the soil significantly alters the
14 arbuscular mycorrhizal species composition and richness and markedly decreases the overall
15 diversity of the arbuscular mycorrhizal community. The decline in coastal sage scrub species
16 can, therefore, directly be linked to the decline of the arbuscular mycorrhizal community
17 (Edgerton-Warburton and Allen, 2000).
18 In addition to excess nitrogen deposition effects on terrestrial ecosystems of the types noted
19 above (e.g., dominant species shifts and other biodiversity impacts), direct atmospheric nitrogen
20 deposition and increased nitrogen inputs via runoff into streams, rivers, lakes, and oceans can
21 have notable impacts on aquatic ecosystems as well. One illustrative example is recently
22 reported research (Paerl et al., 2001) characterizing impacts of nitrogen deposition on the
23 Pamlico Sound, NC, estuarine complex, which serves as a key fisheries nursery supporting an
24 estimated 80% of commercial and recreational fmfish and shellfish catches in the southeastern
25 U.S. Atlantic coastal region. Such direct atmospheric nitrogen deposition onto waterways
26 feeding into the Pamlico Sound or onto the sound itself and indirect nitrogen inputs via runoff
27 from upstream watersheds contribute to conditions of severe water oxygen depletion; formation
28 of algae blooms in portions of the Pamlico Sound estuarine complex; altered fish distributions,
29 catches, and physiological states; and the incidence of disease. Under extreme conditions of
30 especially high rainfall rate events (e.g., hurricanes) affecting watershed areas feeding into the
31 sound, the effects of nitrogen runoff (in combination with excess loadings of metals or other
April 2002 4-96 DRAFT-DO NOT QUOTE OR CITE
-------
1 nutrients) can be massive—e.g., creation of the widespread "dead-zone" affecting large areas of
2 the Pamlico Sound for many months after hurricane Fran in 1996 and hurricanes Dennis, Floyd,
3 and Irene in 1999 impacted eastern North Carolina.
4 Nitrogen saturation of a high elevation watershed in the southern Appalachian Mountains
5 was observed to affect streamwater chemistry. The Great Smoky Mountains in the southeastern
6 United States receive high total atmospheric deposition of sulfur and nitrogen (2,200 Eq/ha/yr of
7 total sulfur and approximately 1,990 Eq/ha/yr of total nitrogen). A major portion of the
8 atmospheric loading is from dry and cloud deposition. Extensive surveys conducted in October
9 1993 and March 1994 indicated that stream pH values were near or below pH 5.5 and that the
10 Acid Neutralizing Capacity (ANC) was below 50 //eq/L at high elevations. Analysis of
11 streamwater indicated that nitrate was the dominant anion (Flum and Nodvin, 1995; Nodvin et
12 al., 1995). The study was expanded to the watershed scale with monitoring of precipitation,
13 thoughfall, stream hydrology, and stream chemistry. Nitrogen saturation of the watershed
14 resulted in extremely high exports of nitrate and promoted both chronic and episodic stream
15 acidification in which the nitrate was the dominant ion. Significant exports of base cation was
16 also observed. Nitrification of the watershed soils resulted in elevations of soil solution
17 aluminum concentrations to levels known to inhibit calcium uptake in red spruce (Nodvin et al.,
18 1995).
19 In the Northeast, nitrogen is the element most responsible for eutrophication in coastal
20 waters of the region (Jaworski et al., 1997). There has been a 3 to 8-fold increase in nitrogen
21 flux from 10 watersheds in the Northeastern United States since the early 1900s. These increases
22 are associated with nitrogen oxide emissions from combustion which have increased 5-fold.
23 Riverine nitrogen fluxes have been correlated with atmospheric deposition onto their landscapes
24 and also with nitrogen oxides emissions into their airsheds. Data from 10 benchmark watersheds
25 with good historical records, indicate that ca. 36-80% of the riverine total nitrogen export, with
26 an average of 64%, was derived directly or indirectly from nitrogen oxide emissions (Jaworski
27 etal., 1997).
28 Excessive nitrogen loss is a symptom of terrestrial ecosystem dysfunction and results in the
29 degradation of water quality and potentially deleterious effects on terrestrial and aquatic
30 ecosystems (Fenn and Poth, 1999). Data from a number of hydrologic, edaphic, and plant
31 indicators indicate that the mixed conifer forests and chaparral systems directly exposed to air
April 2002 4-97 DRAFT-DO NOT QUOTE OR CITE
-------
1 pollution from greater Los Angeles are nitrogen saturated. Preliminary data suggests that
2 symptoms of nitrogen saturation are evident in mixed conifer or chaparral sites receiving
3 atmospheric deposition of 20 to 25 kg/N/ha/y (Fenn et al, 1996). Available data clearly indicate
4 that ecosystems with a Mediterranean climate have a limited capacity to retain nitrogen within
5 the terrestrial system (Fenn and Poth, 1999). A 3-year study of streamwater NO3" concentrations
6 along nitrogen deposition gradients in the San Bernardino Mountains in southern California
7 evaluated streamwater quality and whether the streamwater concentrations covaried with
8 nitrogen deposition across pollution gradients in the San Bernardino Mountains. Streamwater
9 NO3" concentrations at Devil Canyon in the San Gabriel Mountains northeast of Los Angeles are
10 the highest reported in North America for forested watersheds (Fenn and Poth, 1999). Five of
11 the six streams monitored maintained elevated NO3" throughout the year. Peak nitrate
12 concentrations ranged from 40 to 350 //mol/L. In the San Gorgonio Wilderness, an area of low
13 to moderate deposition where 12 streams were sampled, only the five that had the greatest air
14 pollution exposure had high NO3" concentrations. The results of the study suggested a strong
15 association between levels of NO3" export in streamwater and the severity of chronic nitrogen
16 deposition to the terrestrial watersheds. However, nitrogen processing within terrestrial and
17 aquatic systems, even in areas with high nitrogen deposition, determine streamwater NO3"
18 concentrations (Fenn and Poth, 1999). The Fernow Experimental Forest in West Virginia, the
19 Great Smoky Mountains National park in Tennessee, and watersheds in southwestern
20 Pennsylvania are the only undisturbed forested sites in North America known to have
21 streamwater NO3" concentrations within the range of values found at Devil Canyon (Fenn and
22 Poth, 1999).
23
24 Effects of Sulfur Deposition. Sulfur is an essential plant nutrient and, as such, is a major
25 component of plant proteins. The most important source of sulfur is sulfate taken up from the
26 soil by plant roots even though plants can utilize atmospheric SO2 (Marschner, 1995). The
27 availability of organically bound sulfur in soils depends largely on microbial decomposition, a
28 relatively slow process. The major factor controlling the movement of sulfur from the soil into
29 vegetation is the rate of release from the organic to the inorganic compartment (May et al., 1972;
30 U.S. Environmental Protection Agency, 1982; Marschner, 1995). Sulfur plays a critical role in
31 agriculture as an essential component of the balanced fertilizers needed to grow and increase
April 2002 4-98 DRAFT-DO NOT QUOTE OR CITE
-------
1 worldwide food production (Ceccotti and Messick, 1997). Atmospheric deposition is an
2 important component of the sulfur cycle. This is true not only in polluted areas where
3 atmospheric deposition is very high, but also in areas of low sulfur input. Additions of sulfur
4 into the soil in the form of SO4"2 could alter the important organic-sulfur/organic-nitrogen
5 relationship involved in protein formation in plants. The biochemical relationship between sulfur
6 and nitrogen in plant proteins and the regulatory coupling of sulfur and nitrogen metabolism
7 indicate that neither element can be assessed adequately without reference to the other. Sulfur
8 deficiency reduces nitrate reductase and, to a similar extent, also glutamine synthetase activity.
9 Nitrogen uptake in forests, therefore, may be loosely regulated by sulfur availability, but sulfate
10 additions in excess of needs do not necessarily lead to injury (Turner and Lambert, 1980; Hogan
11 etal., 1998).
12 Only two decades ago, there was little information comparing sulfur cycling in forests with
13 other nutrients, especially nitrogen. With the discovery of deficiencies in some unpolluted
14 regions (Kelly and Lambert, 1972; Humphreys et al., 1975; Turner et al., 1977; Schnug, 1997)
15 and excesses associated with acidic deposition in other regions of the world (Meiwes and
16 Khanna, 1981; Shriner and Henderson, 1978; Johnson et al., 1982a,b), interest in sulfur nutrition
17 and cycling in forests has heightened. General reviews of sulfur cycling in forests have been
18 written by Turner and Lambert (1980), Johnson (1984), Mitchell et al. (1992a,b), and Hogan
19 et al. (1998). The salient elements of the sulfur cycle as it may be affected by changing
20 atmospheric deposition are summarized by Johnson and Mitchell (1998). Sulfur has become the
21 most important limiting factor in European agriculture because of the desulfurization of
22 industrial emissions (Schnug, 1997).
23 Most studies dealing with the impacts of sulfur deposition on plant communities have been
24 conducted in the vicinity of point sources and have investigated above-ground effects of SO2 or
25 acidifying effects of sulfate on soils (Krupa and Legge, 1998; Dreisinger and McGovern, 1970;
26 Legge, 1980; Winner and Bewley, 1978a,b; Laurenroth and Michunas, 1985; U.S. Environmental
27 Protection Agency, 1982). Krupa and Legge (1986), however, observed a pronounced increase
28 in foliar sulfur concentrations in all age classes of needles of the hybrid pine lodgepole x jack
29 pine (Pinus contorta x P. banksiand). This vegetation had been exposed to chronic low
30 concentrations of sulfur dioxide (SO2) and hydrogen sulfide (H2S) for more than 20 years and,
31 then, to fugitive sulfur aerosol. Observations under the microscope showed no sulfur deposits on
April 2002 4-99 DRAFT-DO NOT QUOTE OR CITE
-------
1 the needle surfaces and led to the conclusion that the sulfur in the needles was derived from the
2 soil. The oxidation of elemental sulfur and the generation of protons is well known for the soils
3 of Alberta, Canada. This process is mediated by bacteria of the Thiobacillus sp. As elemental
4 sulfur gradually is converted to protonated SO4, it can be leached downward and readily taken up
5 by plant roots. The activity of Thiobacillus sp. is stimulated by elemental sulfur additions (Krupa
6 and Legge, 1986).
7
8 Effects of Acidic Deposition on Forest Soils. Acidic deposition over the past quarter of a
9 century has emerged as a critical environmental stress that affects forested landscapes and aquatic
10 ecosystems in North America, Europe, and Asia (Driscoll et al., 2001). Acidic deposition can
11 originate from transboundary air pollution and affect large geographic areas. It is composed of
12 ions, gases, particles derived from gaseous emissions of sulfur dioxide (SO2), nitrogen oxides
13 (NOX), ammonia (NH3), and particulate emissions of acidifying and neutralizing compounds and
14 is highly variable across space and time. It links air pollution to diverse terrestrial and aquatic
15 ecosystems and alters the interactions of the hydrogen ion (H+) and many elements (e.g., sulfur,
16 nitrogen, calcium, magnesium, and aluminum). Acidic deposition contributes directly and
17 indirectly to biological stress and the degradation of ecosystems and has played a major role in
18 recent acidification of soil in some areas of Europe and, to a more limited extent, eastern North
19 America (Driscoll et al., 2001).
20 Substantial and previously unsuspected changes in soils have been observed in polluted
21 areas of eastern North America, the United Kingdom, Sweden, and Central Europe and in less
22 polluted regions of Australia and western North American (reviewed by Johnson et al., 1999 and
23 by Huntington, 2000). In some cases, trends were toward more acidic soils (e.g., Markewitz
24 et al., 1998), and, in others, there were no consistent trends, with some soils showing increases
25 and some showing decreases at different sampling times, and some showing no change (e.g.,
26 Johnson and Todd, 1998; Trettin et al., 1999; Yanai et al., 1999).
27 Significant changes in soil chemistry have occurred at many sites in the eastern United
28 States during recent decades. Patterns of change in tree ring chemistry, principally at high
29 elevations sites in the eastern United States, reflect the changing inputs of regional pollutants to
30 forests. A temporal sequence of changes in uptake patterns, and possibly in tree growth, would
31 be expected if significant base cation mobilization and depletion of base cations from eastern
April 2002 4-100 DRAFT-DO NOT QUOTE OR CITE
-------
1 forest soils has occurred. Temporal changes in the chemistry of tree rings of red spruce were
2 examined as indicators of historical changes in the chemical environment of red spruce.
3 Analysis of changes in wood chemistry from samples across several sites indicated that
4 there have been substantial departures from the expected linear decreases in calcium
5 accumulation patterns in wood. A region-wide calcium increase above expected levels followed
6 by decreasing changes in wood calcium suggest that calcium mobilization began possibly 30 to
7 40 years ago and has been followed by reduced accumulation rates in wood, presumably
8 associated with decreasing calcium availability in soil (Bondietti and McLaughlin, 1992). The
9 period of calcium mobilization coincides with a region-wide increase in growth rate of red
10 spruce; whereas the period of decreasing levels of calcium in wood corresponds temporally with
11 patterns of decreasing radial growth at high elevation sites throughout the region during the past
12 20 to 30 years. The decline in wood calcium suggests that calcium loss may have increased to
13 the point at which base saturation of soils has been reduced. Increases in aluminum and iron
14 typically occur as base cations are removed from the soils by tree uptake (Bondietti and
15 McLaughlin, 1992). The changes are spatially and temporally consistent with changes in the
16 emissions of SO2 and NO2 across the region and suggest that increased acidification of soils has
17 occurred.
18 Studies by Shortle and Bondietti (1992) support the view that changes in soil chemistry in
19 eastern North America forest sites occurred many decades ago, "before anybody was looking."
20 Sulfur and nitrogen emissions began increasing in eastern North America in the 1920s and
21 continued to increase into the 1980s when sulfur began to decrease but nitrogen emissions did
22 not (Garner et al., 1989). Shortle and Bondietti (1992) present evidence that, from the late 1940s
23 into the 1960s, the mor humus (organic) layer of acid-sensitive forest sites in eastern North
24 America underwent a significant change that resulted in the loss of exchangeable essential base
25 cations and interrupted the critical base nutrient cycles between mature trees and the root-humus
26 complex. The timing of the effect appears to have coincided with the period when the SOX and
27 NOX emissions in eastern North America subject to long-range transport were increasing the most
28 rapidly (see above; Shortle and Bondietti, 1992). Although forest ecosystems other than the
29 high-elevation spruce-fir forests are not currently manifesting symptoms of injury directly
30 attributable to acid deposition, less sensitive forests throughout the United States are
31 experiencing gradual losses of base cation nutrients, which in many cases will reduce the quality
April 2002 4-101 DRAFT-DO NOT QUOTE OR CITE
-------
1 of forest nutrition over the long term (National Science and Technology Council, 1998). In some
2 cases it may not even take decades because these forests already have been receiving sulfur and
3 nitrogen deposition for many years. The current status of forest ecosystems in different U.S.
4 geographic regions varies, as does their sensitivity to nitrogen and sulfur deposition. Variation in
5 potential future forest responses or sensitivity are caused, in part, by differences in deposition of
6 sulfur and nitrogen, ecosystem sensitivities to sulfur and nitrogen additions, and responses of
7 soils to sulfur and nitrogen inputs (National Science and Technology Council, 1998).
8 Acidic deposition has played a major role in recent soil acidification in some areas of
9 Europe and, to a more limited extent, eastern North America. Examples include the study by
10 Hauhs (1989) at Lange Bramke, Germany, which indicated that leaching was of major
11 importance in causing substantial reduction in soil-exchangeable base cations over a 10-year
12 period (1974-1984). Soil acidification and its effects result from the deposition of nitrate (NO3")
13 and sulfate (SO4"2) and the associated hydrogen (H+) ion. The effects of excessive nitrogen
14 deposition on soil acidification and nutrient imbalances have been well established in Dutch
15 forests (Van Breemen et al., 1982; Roelofs et al., 1985; Van Dijk and Roelofs, 1988).
16 For example, Roelofs et al. (1987) proposed that NH3 /NH4+ deposition leads to heathland
17 changes via two modes: acidification of the soil and the loss of cations K+, Ca+2, and Mg+2; and
18 nitrogen enrichment that results in "abnormal" plant growth rates and altered competitive
19 relationships. Nihlgard (1985) suggested that excessive nitrogen deposition may contribute to
20 forest decline in other specific regions of Europe. Falkengren-Grerup (1987) noted that, during
21 about 50 years, unexpectedly large increases in growth of beech (Fagus sylvatica L.) were
22 associated with decreases in pH and exchangeable cations in some sites in southernmost Sweden.
23 Likens et al. (1996, 1998) suggested that soils are changing at the Hubbard Brook
24 Watershed, NH, because of a combination of acidic deposition and reduced base cation
25 deposition. They surmised, based on long-term trends in streamwater data, that large amounts of
26 calcium and magnesium have been lost from the soil-exchange complex over a 30-year period
27 from approximately 1960 to 1990. The authors speculate that the declines in base cations in soils
28 may be the cause of recent slowdowns in forest growth at Hubbard Brook. In a follow-up study,
29 however, Yanai et al. (1999) found no significant decline in calcium and magnesium
30 concentrations in forest floors at Hubbard Brook over the period 1976 to 1997. They also found
31 both gains and losses in forest floor calcium and magnesium between 1980 and 1990 in a
April 2002 4-102 DRAFT-DO NOT QUOTE OR CITE
-------
1 regional survey. Thus, they concluded that "forest floors in the region are not currently
2 experiencing rapid losses of base cations, although losses may have preceded the onset of these
3 three studies." The biogeochemistry of calcium at Hubbard Brook is discussed in detail by
4 Likens etal. (1998).
5 Hydrogen ions entering a forest ecosystem first encounter the forest canopy, where they are
6 often exchanged for base cations that then appear in throughfall (Figure 4-10 depicts a model of
7 FT sources and sinks). Base cations leached from the foliage must be replaced through uptake
8 from the soil, or foliage cations will be reduced by the amounts leached. In the former case, the
9 acidification effect is transferred to the soil where H + is exchanged for a base cation at the
10 root-soil interface. Uptake of base cations or NH4 + by vegetation or soil microorganisms causes
11 the release of H + in order to maintain charge balance; uptake of nutrients in anionic form (NO3",
12 SO4"2, PO4"3) causes the release of OH "in order to maintain charge balance. Thus, the net
13 acidifying effect of uptake is the difference between cation and anion uptake. The form of ions
14 taken up is known for all nutrients but nitrogen because either NH4+ or NO3" can be utilized.
15 In that nitrogen is a nutrient taken up in greatest quantities, the uncertainty in the ionic form of
16 nitrogen utilized creates great uncertainty in the overall H+ budget for soils (Johnson 1992).
17 The cycles of base cations differ from those of N, P, and S in several respects. The fact that
18 calcium, potassium, and magnesium exist primarily as cations in solution, whereas N, P, and
19 S exist primarily as anions, has major implications for the cycling of the nutrients and the effects
20 of acid deposition on these cycles. The most commonly accepted model of base cation cycling in
21 soils is one in which base cations are released by weathering of primary minerals to cation
22 exchange sites where they are available for either plant uptake or leaching (Figure 4-10). The
23 introduction of H + by atmospheric deposition or by internal processes will affect the fluxes of
24 Ca, K, and Mg via cation exchange or weathering processes. Therefore, soil leaching is often of
25 major importance in cation cycles, and many forest ecosystems show a net loss of base cations
26 (Johnson, 1992).
27 Two basic types of soil change are involved: (1) a short-term intensity type change
28 resulting from the concentrations of chemicals in soil water and (2) a long-term capacity change
29 based on the total content of bases, aluminum, and iron stored in the soil (Reuss and Johnson,
30 1986; Van Breemen et al., 1983). Changes in intensity factors can have a rapid affect on the
31 chemistry of soil solutions. Increases in the amounts of sulfur and nitrogen in acidic deposition
April 2002 4-103 DRAFT-DO NOT QUOTE OR CITE
-------
Deposition
H+
Soil
Organism
Uptake
Uptake
Tree
2H+ + NO3
Nitrification
CO2 + H2O
Carbonic Acid Formation
R-COOH
Organic Acid Formation
Anion
Adsorption
SOf
2OH"
Soil
Organism
Uptake
2OH
Leaching
Figure 4-10. Schematic of sources and sinks of hydrogen ions in a forest (from Taylor
et al., 1994).
1 can cause immediate increases in acidity and mobilization of aluminum in soil solutions.
2 Increased aluminum concentrations and an increase in the Ca/Al ratio in soil solution have been
3 linked to a significant reduction in the availability of essential base cations to plants, an increase
4 in plant respiration, and increased biochemical stress (National Science and Technology Council,
5 1998).
6 Rapid changes in intensity resulting from the addition of increased amounts of nitrogen or
7 sulfur in acidic deposition can have a rapid effect on the chemistry of soil solutions by increasing
8 the acidity and mobilizing aluminum. Increased concentrations of aluminum and an increase in
9 the ratio of calcium to aluminum in soil solution have been linked to a significantly reduced
10 availability of essential cations to plants.
April 2002
4-104
DRAFT-DO NOT QUOTE OR CITE
-------
1 Capacity changes are the result of many factors acting over long time periods. The content
2 of base cations (calcium, magnesium, sodium, and potassium) in soils results from additions
3 from the atmospheric deposition, decomposition of vegetation, and geologic weathering. Loss of
4 base cations may occur through plant uptake and leaching. Increased leaching of base cations
5 may result in nutrient deficiencies in soils as has been happening in some sensitive forest
6 ecosystems (National Science and Technology Council, 1998).
7 Aluminum toxicity is a possibility in acidified soils. Atmospheric deposition (or any other
8 source of mineral anions) can increase the concentration of Al, especially A13+, in soil solution
9 without causing significant soil acidification (Johnson and Taylor, 1989). Aluminum can be
10 brought into soil solution in two ways: (1) by acidification of the soil and (2) by an increase in
11 the total anion and cation concentration of the soil solution. The introduction of mobile, mineral
12 acid anions to an acid soil will cause increases in the concentration of aluminum in the soil
13 solution, but extremely acid soils in the absence of mineral acid anions will not produce a
14 solution high in aluminum. An excellent review of the relationships among the most widely used
15 cation-exchange equations and their implications for the mobilization of aluminum into soil
16 solution is provided by Reuss (1983).
17 A major concern has been that soil acidity would lead to nutrient deficiency. Calcium is
18 essential for root development and the formation of wood, and it plays a major role in cell
19 membrane integrity and cell wall structure. Aluminum concentrations in the soil can influence
20 forest tree growth in regions where acidic deposition and natural acidifying processes increase
21 soil acidity. Acidic deposition mobilizes calcium and magnesium, which are essential for root
22 development and stem growth. Mobilized aluminum can also bind to fine root tips of red spruce,
23 further limiting calcium and magnesium uptake (Shortle and Smith, 1988; Shortle et al., 1997).
24 There is abundant evidence that aluminum is toxic to plants. Upon entering tree roots, it
25 accumulates in root tissues (Thornton et al., 1987; Vogt et al., 1987a, b). Reductions in calcium
26 uptake have been associated with increases in aluminum uptake (Clarkson and Sanderson, 1971).
27 A number of studies suggest that the toxic effect of aluminum on forest trees could be caused by
28 Ca+2 deficiency (Shortle and Smith, 1988; Smith, 1990a). Mature trees have a high calcium
29 requirement relative to agriculture crops (Rennie, 1955). Shortle and Smith (1988) attributed the
30 decline of red spruce in eight stands across northern New England from Vermont to Maine to an
31 imbalance of A13+ and Ca+2 in fine root development.
April 2002 4-105 DRAFT-DO NOT QUOTE OR CITE
-------
1 To be taken up from the soil by roots, calcium must be dissolved in soil water (Lawrence
2 and Huntingdon, 1999). Aluminum in soil solution reduces calcium uptake by competing for
3 binding sites in the cortex of fine roots. Tree species may be adversely affected if high aluminum
4 to nutrient ratios create a nutrient deficiency by limiting uptake of calcium and magnesium
5 (Shortle and Smith, 1988; Garner, 1994). Acid deposition, by lowering the pH of aluminum-rich
6 soil, can increase aluminum concentrations in soil water through dissolution and ion exchange
7 processes. Aluminum is more readily taken up than is calcium because of its greater affinity for
8 negatively charged surfaces. When present in the forest floor, aluminum tends to displace
9 adsorbed calcium and causes it to be more readily leached. The continued buildup of aluminum
10 in the forest floor layer, where nutrient uptake is greatest, can lower efficiency of calcium uptake
11 when the ratio of calcium to aluminum in soil water is less than one (Lawrence and Huntington,
12 1999). Reduction in calcium uptake suppresses cambial growth and reduces the rate of wood
13 (annual ring) formation, decreases the amount of functional sapwood and live crown, and
14 predisposes trees to disease and injury from stress agents when the functional sapwood becomes
15 less that 25% of cross-sectional stem area (Smith, 1990a). A 1968 Swedish report to the United
16 Nations postulated a decrease in forest growth of ca. 1.5% per year when the ratio of calcium to
17 aluminum in soil water is less than one (Lawrence and Huntington, 1999). The concern that
18 acidification and nutrient deficiency may result in forest decline remains today.
19 Acidic deposition has been firmly implicated as a causal factor in northeastern high-
20 elevation decline of red spruce (DeHayes et al., 1999). The frequency of freezing injury of red
21 spruce has increased over the past 40 years, a period that coincides with increase emissions of
22 sulfur and nitrogen oxides and acidic deposition (DeHayes et al., 1999). Studies indicate that
23 there is a significant positive association between cold tolerance and foliar calcium in trees
24 exhibiting deficiency in foliar calcium. Most of the calcium in conifer needles is insoluble
25 calcium oxalate and pectate crystals, which are of little physiological importance. It is the labile
26 calcium ions in equilibrium within the plasma membrane that are of major physiological
27 importance (DeHayes et al., 1999). The membrane-associated pool of calcium (mCa), although a
28 relatively small fraction of total foliar ion pools, strongly influences the response of cells to
29 changing environmental conditions. The plant plasma membrane plays a critical role in
30 mediating cold acclimation and low-temperature injury. Leaching of calcium associated with
31 acidic deposition is considered to be the result of cation exchange due to exposure to the H+ ion.
April 2002 4-106 DRAFT-DO NOT QUOTE OR CITE
-------
1 The studies of DeHayes et al. (1999) demonstrate that the direct deposition of acidic deposition
2 on needles represents a unique environmental stress, in that it preferentially removes mCa which
3 is not readily replaced in autumn. They propose that direct deposition on red spruce foliage
4 preferentially displaces calcium ions specifically associated with plasma membranes of
5 mesophyll cells resulting in the reduction of mCa and the destabilizing of plasma membranes and
6 depletion of messenger calcium. Further, DeHayes et al.(1999) state that their studies raise the
7 strong possibility that acid rain alteration of the mCa and membrane integrity is not unique to red
8 spruce but has been demonstrated in many other northern temperate forest tree species including
9 yellow birch (Betula alleghaniensis), white spruce (Picea glaucus), red maple (Acer rubruni)
10 eastern white pine (Pinus strobus), and sugar maple (Acer saccharuni). Assessments of mCa,
11 membrane integrity, and the effects of other secondary stresses have not yet been made for these
12 species.
13 Seasonal and episodic acidification of surface waters have been observed in the eastern
14 United States, Canada and Europe (Hyer et al., 1995). In the Northeast, the Shenandoah National
15 Park in Virginia, and the Great Smoky Mountains, episodic acidification has been associated
16 with the nitrate ion (Driscoll et al., 2001; Hyer et al., 1995 ; Eshleman et al., 1995). The short-
17 term acid episodes occur during spring snowmelts and large precipitation events (Driscoll et al.,
18 2001). Episodic acidification of surface waters has usually been considered to be a transient loss
19 of acid neutralizing capacity associated with snowmelt/rainfall runoff and, as such, represents
20 short-term (hours to weeks) effects considered to be distinguishable from chronic long-term
21 (years to centuries) changes in acidity. Studies of both episodic and chronic acidification of
22 surface waters indicate that acidification can have long-term adverse effects on fish populations,
23 declines of species richness, abundance of zooplankton, and macroinvertebrates (Driscoll et al.,
24 2001; Eshleman et al., 1995). Nitrogen saturation of soils and the slow release of nitrates has
25 inhibited the recovery of acid sensitive systems (Driscoll et al., 2001). The acidification of
26 aquatic ecosystems and the effects on aquatic biota has been discussed in greater detail in the Air
27 Quality Criteria for Nitrogen Oxides (U.S. Environmental Protection Agency, 1993).
28 Air pollution is not the sole cause of soil change. High rates of acidification are occurring
29 in less polluted regions of the western United States and Australia because of internal soil
30 processes, such as tree uptake of nitrate and nitrification associated with excessive nitrogen
31 fixation (Johnson et al., 1991b). Many studies have shown that acidic deposition is not a
April 2002 4-107 DRAFT-DO NOT QUOTE OR CITE
-------
1 necessary condition for the presence of extremely acid soils, as evidenced by their presence in
2 unpolluted, even pristine, forests of the northwestern United States and Alaska (Johnson et al.,
3 1991b). Soil can become acidic when H+ ions attached to NH4+ or HNO3 remain in the soil after
4 nitrogen is taken up by plants. For example, Johnson et al. (1982b) found significant reductions
5 in exchangeable K + over a period of only 14 years in a relatively unpolluted Douglas fir
6 Integrated Forest Study (IFS) site in the Washington Cascades. The effects of acid deposition at
7 this site were negligible relative to the effects of natural leaching (primarily carbonic acid) and
8 nitrogen tree uptake (Cole and Johnson, 1977). Even in polluted regions, numerous studies have
9 shown the importance of tree uptake of NH4+ and NO3" in soil acidification. Binkley et al. (1989)
10 attributed the marked acidification (pH decline of 0.3 to 0.8 units and base saturation declines of
11 30 to 80%) of abandoned agricultural soil in South Carolina over a 20-year period to NH4+ and
12 NO3" uptake by a loblolly pine plantation.
13 An interesting example of uptake effects on soil acidification is that of Al uptake and
14 cycling (Johnson et al., 1991b). Aluminum accumulation in the leaves of coachwood
15 (Ceratopetalum apetalum) in Australia has been found to have a major effect on the distribution
16 and cycling of base cations (Turner and Kelly, 1981). The presence of C. apetalum as a
17 secondary tree layer beneath brush box (Lophostemon confertus) was found to lead to increased
18 soil exchangeable Al3+ and decreased soil exchangeable Ca2+ (Turner and Kelly, 1981). The
19 constant addition of aluminum-rich litter fall obviously has had a substantial effect on soil
20 acidification, even if base cation uptake is not involved directly.
21 Given the potential importance of particulate deposition for base cation status of forest
22 ecosystems, the findings of Driscoll et al. (1989, 2001) and Hedin et al. (1994) are especially
23 relevant. Driscoll et al. (1989, 2001) noted a decline in both SO4"2 and base cations in both
24 atmospheric deposition and stream water over the past two decades at Hubbard Brook
25 Watershed, NH. The decline in SO4"2 deposition was attributed to a decline in emissions, and the
26 decline in stream water SO4"2 was attributed to the decline in sulfur deposition.
27 Hedin et al. (1994) reported a steep decline in atmospheric base cation concentrations in
28 both Europe and North America over the past 10 to 20 years. The reductions in SO2 emissions in
29 Europe and North America in recent years have not been accompanied by equivalent declines in
30 net acidity related to sulfate in precipitation. These current declines in sulfur deposition have, in
31 varying degrees, been offset by declines in base cations and may be contributing "to the increased
April 2002 4-108 DRAFT-DO NOT QUOTE OR CITE
-------
1 sensitivity of poorly buffered systems." Analysis of the data from the IPS supports the authors'
2 contention that atmospheric base cation inputs may seriously affect ecosystem processes.
3 Johnson et al. (1994b) analyzed base cation cycles at the Whiteface Mountain IPS site in detail
4 and concluded that Ca losses from the forest floor were much greater than historical losses, based
5 on historical changes in forest floor Ca observed in an earlier study. Further, the authors suggest
6 that the difference between historical and current net loss rates of forest floor Ca may be caused
7 by sharply reduced atmospheric inputs of calcium after about 1970 and may be exacerbated by
8 sulfate leaching (Johnson et al., 1994b).
9 The calcium/aluminum molar ratio has been suggested as a valuable ecological indicator of
10 an approximate threshold beyond which the risk of forest injury from Al stress and nutrient
11 imbalances increases (Cronan and Grigal, 1995). The Ca/Al ratio also can be used as an
12 indicator to assess forest ecosystem changes over time in response to acidic deposition, forest
13 harvesting, or other process that contribute to acid soil infertility. This ratio, however, may not
14 be a reliable indicator of stress in areas with both high atmospheric deposition of ammonium and
15 magnesium deficiency via antagonism involving ammonium rather than aluminum and in areas
16 with soil solutions with calcium concentrations greater than 500 micromoles per liter (National
17 Science and Technology Council, 1998). Cronan and Grigal (1995), based on a review of the
18 literature, have made the following estimates for determining the adverse impact of acidic
19 deposition on tree growth or nutrition:
20 • forests have a 50% risk of adverse impacts if the Ca/Al ration is 1.0,
21 • the risk is 75% if the ratio is 0.5, and
22 • the risk approaches 100% if the ratio is 0.2.
23 The Ca/Al ratio of soil solution provides only an index of the potential for Al stress. Cronan and
24 Grigal (1995) state that the overall uncertainty of the Ca/Al ratio associated with a given
25 probability ratio is considered to be approximately ±50%. Determination of thresholds for
26 potential forest impacts requires the use of the four successive measurement endpoints in the soil,
27 soil solution, and plant tissue listed below.
28 (1) Soil base saturation less than 15% of effective cation exchange capacity,
29 (2) Soil solution Ca/Al molar ratio less than 1.0 for 50% risk,
30 (3) Fine roots tissue Ca/Al molar ratio less than 0.2 for 50% risk, and
31 (4) Foliar tissue Ca/Al molar ratio less than 12.5 for 50% risk.
April 2002 4-109 DRAFT-DO NOT QUOTE OR CITE
-------
1 The application of the Ca/Al ratio indicator for assessment and monitoring of forest health risks
2 has been recommended for sites or in geographic regions where the soil base saturation <15%.
3
4 Critical Loads. In Europe, the critical load concept generally has been accepted as the
5 basis for abatement strategies to reduce or prevent injury to the functioning and vitality of forest
6 ecosystems caused by long-range transboundary acidic deposition (Lokke, et al., 1996). The
7 critical load has been defined as a "quantitative estimate of an exposure to one or more pollutants
8 below which significant harmful effects on specified sensitive elements of the environment do
9 not occur according to present knowledge" (Lokke et al., 1996). A biological indicator, a
10 chemical criterion, and a critical value are the elements used in the critical load concept. The
11 biological indicator is the organism used to indicate the status of the receptor ecosystem; the
12 chemical criterion is the parameter that results in harm to the biological indicator; and the critical
13 value is the value of the chemical criterion below which no significant harmful response occurs
14 to the biological indicator (Lokke et al., 1996). Trees, and sometimes other plants, are used as
15 the biological indicators in the case of critical loads for forests. The critical load calculation
16 using the current methodology, is essentially an acidity/alkalinity mass balance calculation. The
17 chemical criterion must be expressible in terms of alkalinity. Initially, the Ca/Al ratio was used;
18 but, recently, the (Ca+Mg+K)/Al ratio has been used (Lokke et al., 1996).
19 Ideally, changes in acidic deposition should result in changes in the status of the biological
20 indicator used in the critical load calculation. However, the biological indicator is the integrated
21 response to a number of different stresses. Furthermore, there are other organisms more sensitive
22 to acid deposition than trees. At high concentrations, Al3+ is known to be toxic to plants,
23 inhibiting root growth and, ultimately, plant growth and performance (Lokke et al., 1996;
24 National Science and Technology Council, 1998). Sensitivity to Al varies considerably between
25 species and within species because of changes in nutritional demands and physiological status
26 that are related to age and climate. Experiments have shown that there are large variations in Al
27 sensitivity, even among ecotypes.
28 Mycorrhizal fungi as possible biological indicators have been suggested by Lokke et al.
29 (1996) because they are intimately associated with tree roots, depend on plant assimilates, and
30 play an essential role in plant nutrient uptake influencing the ability of their host plants to tolerate
31 different anthropogenically generated stresses. Mycorrhizas and fine roots are an extremely
April 2002 4-110 DRAFT-DO NOT QUOTE OR CITE
-------
1 dynamic component of below-ground ecosystems and can respond rapidly to stress. They have a
2 relatively short life span, and their turnover appears to be strongly controlled by environmental
3 factors. Changes in mycorrhizal species composition or the loss of dominant mycorrhizal species
4 in areas where diversity is already low may lead to increased susceptibility of plant to stress
5 (Lokke et al., 1996). Stress affects the total amount of carbon fixed by plants and modifies
6 carbon allocation to biomass, symbionts, and secondary metabolites. The physiology of carbon
7 allocation has also been suggested as an indicator of anthropogenic stress (Andersen and
8 Rygiewicz, 1991). Because mycorrhizal fungi are dependent for their growth on the supply of
9 assimilates from the host plants, stresses that shift the allocation of carbon reserves to the
10 production of new leaves at the expense of supporting tissues will be reflected rapidly in
11 decreased fine root and mycorrhizzal biomass (Winner and Atkinson, 1986). Decreased carbon
12 allocation to roots also affects soil carbon and rhizosphere organisms. Soil dwelling animals are
13 important for decomposition, soil aeration, and nutrient redistribution in the soil. They
14 contribute to decomposition and nutrient availability mainly by increasing the accessibility of
15 dead plant material to microorganisms. Earthworms decrease in abundance and in species
16 number in acidified soils (Lokke et al., 1996).
17
18 Effects of Wet and Dry Deposition on Biogeochemical Cycling—The Integrated Forest
19 Study. The Integrated Forest Study (IPS) (Johnson and Lindberg, 1992a) has provided the most
20 extensive data set available on wet and dry deposition and deposition effects on the cycling of
21 elements in forest ecosystems. The overall patterns of deposition and cycling have been
22 summarized by Johnson and Lindberg (1992a), and the reader is referred to that reference for
23 details. The following is a summary of particulate deposition, total deposition, and leaching in
24 the IPS sites.
25 Particulate deposition in the IPS was separated at the 2-yam level; a decision was made to
26 include total particulate deposition in this analysis and may include the deposition of particles
27 larger than 10 //m.
28 Particulate deposition contributes considerably to the total impact of base cations to most of
29 the IPS sites. On average, particulate deposition contributes 47% to total calcium deposition
30 (range: 4 to 88%), 49% of total potassium deposition (range: 7 to 77%), 41% to total magnesium
31 deposition (range: 20 to 88%), 36% to total sodium deposition (range: 11 to 63%), and 43% to
April 2002 4-111 DRAFT-DO NOT QUOTE OR CITE
-------
1 total base cation deposition (range: 16 to 62%). Of the total particulate deposition, the vast
2 majority (>90%) is >2 //m.
3 Figures 4-11 through 4-14 summarize the deposition and leaching of calcium, magnesium,
4 potassium, and total base cations for the IPS sites. As noted in the original synthesis (Johnson
5 and Lindberg, 1992a), some sites show net annual gains of base cations (i.e., total deposition
6 > leaching), some show losses (total deposition < leaching), and some are approximately in
7 balance. Not all cations follow the same pattern at each site. For example, calcium shows net
8 accumulation at the Coweeta, TN, Durham (Duke), NC, and Florida sites (Figure 4-11).
9 Potassium shows accumulation at the Duke, Florida, Douglas-fir; red alder, Thompson, WA,
10 Huntingdon Forest, NY, and Whiteface Mountain, NY, sites (Figure 4-13). Magnesium
11 accumulated only at the Florida sites (Figure 4-12); only at the Florida site, is there a clear net
12 accumulation of total base cations (Figure 4-14).
13 As noted previously, the factors affecting net calcium accumulation or loss include the soil-
14 exchangeable cation composition; base cation deposition rate; the total leaching pressure because
15 of atmospheric sulfur and nitrogen inputs, as well as natural (carbonic and organic) acids; and
16 biological demand (especially for potassium). At the Florida site, which has a very cation-poor,
17 sandy soil (derived from marine sand), the combination of all these factors leads to net base
18 cation accumulation from atmospheric deposition (Johnson and Lindberg, 1992a). The site
19 showing the greatest net base cation losses, the red alder stand in Washington state, is one that is
20 under extreme leaching pressure by nitrate produced because of excessive fixation by that species
21 (Van Miegroet and Cole, 1984). In the red spruce site in the Smokies, the combined effects of
22 SO4"2 and NO3" leaching are even greater than in the red alder site (Figure 4-15), but a
23 considerable proportion of the cations leached from this extremely acid soil consist of FT and
24 Al+3 rather than of base cations (Johnson and Lindberg, 1992a). Thus, the red spruce site in the
25 Smokies is approximately in balance with respect to calcium and total base cations, despite the
26 very high leaching pressure at this site (Figures 4-11 and 4-14).
27 The relative importance of particulate base cation deposition varies widely with site and
28 cation and is not always related to the total deposition rate. The proportion of calcium deposition
29 in particulate form ranges from a low of 4% at the Whiteface Mountain site to a high of 88% at
30 the Maine site (Figure 4-11). The proportion of potassium deposition as particles ranges from
31 7% at the Smokies site to 77% at the Coweeta site (Figure 4-13), and the proportion of total base
April 2002 4-112 DRAFT-DO NOT QUOTE OR CITE
-------
1,000
500
CD n
CD u
'CD
~
_CD
CO
-500
= -1,000
LU
-1,500 -
-2,000
63%
41%
^
1
J :; 2|jrn
^ Wet
^ Leaching
Pe
55%
rcent c
65%
—
I
//
I
)f total
40%
•
depos
26%
=
I
tion as
25%
§^5
^
I
I
i
5 partic
49%
R
M
:\es:
58%
$
88%
i
4%
1
i
•2J.
^
I
n
)
CP DL GS LP FS
I Warmer Sites
DF RA NS
FF MS WF ST
-Colder Sites *
Figure 4-11. Calcium deposition in >2-/j,m particles, <2-/j,m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
CP = Pinus strobus, Coweeta, TN; DL = Pinus taeda, Durham (Duke), NC;
GS = Pinus taeda, B. F. Grant Forest, GA; LP = Pinus taeda, Oak Ridge, TN;
FS = Pinus eliottii, Bradford Forest, FL; DF = Psuedotsuga menziesii,
Thompson, WA; RA = Alnus rubra; Thompson WA; NS = Picea abies,
Nordmoen, Norway; HF = northern hardwood, Huntington Forest, NY;
MS = Picea rubens, Howland, ME; WF = Picea rubens, Whiteface Mountain,
NY; and ST = Picea rubens, Clingman's Dome, NC.
1 cation deposition ranges from 16% at the Whiteface site to 62% at the Maine site (Figure 4-14).
2 Overall, particulate deposition at the site in Maine accounted for the greatest proportion of
3 calcium, magnesium, potassium, and base cation deposition (88, 88, 57, and 62%, respectively),
4 even though total deposition was relatively low. At some sites, the relative importance of
5 particulate deposition varies considerably by cation. At the Whiteface Mountain site, particulate
6 deposition accounts for 4, 20, and 40% of calcium, magnesium, and potassium deposition,
7 respectively. At the red spruce site in the Smokies, particulate deposition accounts for 46, 26%,
8 7% of calcium, magnesium, and potassium deposition, respectively.
9
April 2002
4-113
DRAFT-DO NOT QUOTE OR CITE
-------
400 -
03
CD
>.
co 0 -
.c
-(—i
0 ?oo -
03 ~Z(J(J
'5
LU -400 -
-600 -
40%
I
w
m
49%
^
I
> 2 |jm
< 2 pin
Wet
Leaching
p<
44%
9rcent
48%
i
i
i
oftota
36%
y
depo-
27%
^
i
sition a
27%
^
I
I
i
s parti
39%
i
cles:
46%
m
88%
i
20%
I
1
26%
—
i
i
CP DL GS LP FS DF RA NS FF MS WF ST
Warmer Sites ^^ Colder Sites
Figure 4-12. Magnesium deposition in >2-(j,m particles, <2-(j,m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
See Figure 4-11 for site abbreviations.
1
2
3
4
5
6
7
8
9
10
11
12
As indicated in the IPS synthesis, SO4"2 and NO^ leaching often are dominated by
atmospheric sulfur and nitrogen (Johnson and Lindberg, 1992a). The exceptions to this are in
cases where natural nitrogen inputs are high (i.e.,the nitrogen-fixing red alder stand), as are NO3"
leaching rates even though nitrogen deposition is low, and where soils adsorb much of the
atmospherically deposited SO4"2 thus reducing SO4"2 leaching compared to atmospheric sulfur
input.
Sulfate and NO3" leaching have a major effect on cation leaching in many of the IPS sites
(Johnson and Lindberg, 1992a). Figure 4-15 shows the total cation leaching rates of the IPS sites
and the degree to which cation leaching is balanced by SO4"2 + NO3" deposition. The SO4"2 and
NO3" fluxes are subdivided further into that proportion potentially derived from particulate sulfur
and nitrogen deposition (assuming no ecosystem retention, a maximum effect) and other sulfur
and nitrogen sources (wet and gaseous deposition, internal production).
April 2002
4-114
DRAFT-DO NOT QUOTE OR CITE
-------
co
*,
'co
c
_CD
CO
'5
CT
UJ
400'
300-
200-
100'
0'
-100-
-200
-300
-400
-500
77%
s
m
76%
^
1
Q > 2 pm
^ Wet
^ Leaching
p<
61%
n
m
Brcent
48%
.— — -
I
of tota
54%
Pi
depos
40%
H
ition a
40%
1
I
\
s partic
68%
1
L/J
elates
23%
i
' 57%
w
m
40%
W
7%
1
i
%
I
!
%
CP DL GS LP FS
I Warmer Sites
DF
RA NS
FF MS WF ST
- Colder Sites ^
Figure 4-13. Potassium deposition in >2-(j,m particles, <2-(j,m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
See Figure 4-11 for site abbreviations.
1 As noted in the IPS synthesis, total SO4"2 and NO3" inputs account for a large proportion
2 (28 to 88%) of total cation leaching in most sites. The exception is the Georgia loblolly pine site,
3 where there were high rates of HCO3" and Cl" leaching (Johnson and Lindberg, 1992a). The role
4 of particulate sulfur and nitrogen deposition in this leaching is generally very small (<10%),
5 however, even if it is assumed that there is no ecosystem sulfur or nitrogen retention.
6 It was noted previously in this chapter that the contribution of particles to total deposition
7 of nitrogen and sulfur at the US sites is lower than that for base cations. On average, particulate
8 deposition contributes 18% to total nitrogen deposition (range: 1 to 33%) and 17% to total sulfur
9 deposition (range: 1 to 30%). Particulate deposition contributes only a small amount to total H+
10 deposition (average = 1%; range: 0 to 2%). (It should be noted, however, that particulate H+
11 deposition in the > 2 //m fraction was neglected.)
April 2002
4-115
DRAFT-DO NOT QUOTE OR CITE
-------
3,000
2,000
1,000
co
0)
cb
- -1,000
-2,000 -
-3,000 -
-4,000
53%
1
47%
E
1
FJ > 2 |jm
^ Leaching
P
48%
i
ercent
62%
I
i
of tota
49%
^_ ^^
1
i
1 depo
28%
1
E
sition £
28%
1
i
i
i
as part
47%
ft
n
culate
44%
i
s:
62%
%
16%
1
31%
1
I
\
CP DL GS LP FS DF RA NS FF MS WF ST
Warmer Sites ^-^ Colder Sites
Figure 4-14. Base cation deposition in >2-^m particles, <2-^m particles, and wet forms
(upper bars) and leaching (lower bars) in the Integrated Forest Study sites.
See Figure 4-11 for site abbreviation.
1 Based on the IPS data, it appears that the parti culate deposition has a greater effect on base
2 cation inputs to soils than on base cation losses associated with inputs of sulfur, nitrogen, and H+.
3 It cannot be determined what fraction of the mass of these particles are <10 //m, but only a very
4 small fraction is <2 //m. These inputs of base cations have considerable significance, not only to
5 the base cation status of these ecosystems, but also to the potential of incoming precipitation to
6 acidify or alkalize the soils in these ecosystems. As noted above, the potential of precipitation to
7 acidify or alkalize soils depends on the ratio of base cations to H+ in deposition, rather than
8 simply on the inputs of H+ alone. In the case of calcium, the term "lime potential" has been
9 applied to describe this ratio; the principle is the same with respect to magnesium and potassium.
10 Sodium is a rather special case, in that it is a poorly absorbing cation and leaching tends to
11 balance input over a relatively short term.
12 Net balances of base cations tell only part of the story as to potential effects on soils; these
13 net losses or gains must be placed in the perspective of the soil pool size. One way to express
April 2002
4-116
DRAFT-DO NOT QUOTE OR CITE
-------
CD
-------
CD
(D
CD
CT
(D
o
CD
_CD
CO
'5
cr
LU
Soil Exchangeable
(Dep - Leaching)*25
-50,000
DL
GS LP FS
Warmer Sites
DF RA
NS
HF MS WF
Colder Sites
ST
Figure 4-16. Soil exchangeable Ca+2 pools and net annual export of Ca+2 (deposition
minus leaching times 25 years) in the Integrated Forest Study sites.
See Figure 4-11 for site abbreviations.
CD
CD
CD
CD
cr
LU
100,000
80,000
Soil Exchangeable
(Dep - Leaching)*25
-20,000
CP DL
GS LP FS
-Warmer Sites -
RA NS HF MS WF ST
+* Colder Sites
Figure 4-17. Soil exchangeable Mg+2 pools and net annual export of Mg+2
(deposition minus leaching times 25 years) in the Integrated Forest
Study sites. See Figure 4-11 for site abbreviations.
April 2002
4-118
DRAFT-DO NOT QUOTE OR CITE
-------
•\ AC\ nnn
•i on nnn -
CO
/i% 1 nn nnn -
' on nnn .
_C
"c fin nnn .
"CD
• Ar\ nnn -
o-
LU
on nnn .
0 •
on nnn
1L
__i — i
0
CP DL
^^m
^~
^m
—
' '
U_
-a-
_LL
Q Soil Exchange
|(Dep- Leachin
i — i
-0-
r~i
able
g)'2b
-u-
GS LP FS DF RA NS HF MS WF ST
Figure 4-18. Soil exchangeable K2+ pools and net annual export of K2+ (deposition
minus leaching times 25 years) in the Integrated Forest Study sites.
See Figure 4-11 for site abbreviations.
1 these soils (if any) will depend on weathering inputs and vegetation outputs, in addition to
2 deposition and leaching. It is noteworthy that each of the sites listed above as sensitive has a
3 large store of weatherable minerals, whereas many of the other soils, with larger exchangeable
4 cation reserves, have a small store of weatherable minerals (e.g., Coweeta white pine, Duke
5 loblolly pine, Georgia loblolly pine, and Oak Ridge loblolly pine) (Johnson and Lindberg, 1992a;
6 April and Newton, 1992).
7 Base cation inputs are especially important to the Smokies red spruce site because of
8 potential aluminum toxicity and calcium and magnesium deficiencies. Johnson et al. (199la)
9 found that soil solution aluminum concentrations occasionally reached levels found to inhibit
10 calcium uptake and cause changes in root morphology in solution culture studies of red spruce
11 (Raynal et al., 1990). In a follow-up study, Van Miegroet et al. (1993) found a slight but
12 significant growth response to calcium and magnesium fertilizer in red spruce saplings near the
13 Smokies red spruce site. Joslin et al. (1992) reviewed soil and solution characteristics of red
14 spruce in the southern Appalachians, and it appears that the IPS site is rather typical.
April 2002
4-119
DRAFT-DO NOT QUOTE OR CITE
-------
1 Wesselink et al. (1995) reported on the complicated interactions among changing
2 deposition and soils at this site (including repeated sampling of soil exchangeable base cation
3 pools) from 1969 to 1991 and compared these results with those of a simulation model. They
4 identified three basic stages of change in this ecosystem. During Stage 1, there was increased
5 deposition of sulfur and constant deposition of base cations, causing increased base cation
6 leaching and reduced base saturation in the soils. During Stage II, sulfur deposition is reduced,
7 and soil solution sulfate and base cation leaching decline accordingly, but base saturation
8 continues to decrease. During Stage HI, two alternative scenarios are introduced: (a) sulfur
9 deposition continues to decline, whereas base cation deposition says constant; or (b) both sulfur
10 and base cation deposition decline. Under Stage ni-a, sulfate and base cation leaching continue
11 to decline, and base saturation begins to increase as base cations displace exchangeable
12 aluminum and cause it to transfer to the gibbsite pool. Under Stage ni-b, this recovery in base
13 saturation is over-ridden by the reduction in base cation deposition.
14 The IPS project, for the first time, accurately quantifies atmospheric deposition inputs to
15 nutrient cycles using state-of-the-art techniques to measure wet and dry deposition. The principal
16 aim of the project was to determine the effects of atmospheric deposition on nutrient status of a
17 variety of forest ecosystems and to determine if these effects are in any way related to current or
18 potential forest decline. Acidic deposition is having a significant effect on nutrient cycling in
19 most of the forest ecosystems studied in the IPS project. The exceptions were the relatively
20 unpolluted Douglas fir, red alder, and Findley Lakes in Washington state. The nature of the
21 effects, however, varies from one location to another (Johnson, 1992). In all but the relatively
22 unpolluted Washington sites, atmospheric deposition was having a significant, often
23 overwhelming effect on cation leaching from the soils. In general, nutrient budget data from IPS
24 and the literature suggest that the susceptibility of southeastern sites to base cation depletion
25 from soils and the development of cation deficiencies by that mechanism appears to be greater
26 than in northern sites (Johnson, 1992).
27 Atmospheric deposition may have affected significantly the nutrient status of some IPS
28 sites through the mobilization of Al. Soil solution Al levels in the Smokies sites approach and
29 sometimes exceed levels noted to impede cation uptake in solution culture studies. It is therefore
30 possible that the rates of base cation uptake and cycling in these sites have been reduced because
31 of soil solution Al. To the extent that atmospheric deposition has contributed to these elevated
April 2002 4-120 DRAFT-DO NOT QUOTE OR CITE
-------
1 soil solution Al levels, it has likely caused a reduction in base cation uptake and cycling rates at
2 these sites. Nitrate and sulfate are the dominant anions in the Smokies sites, and nitrate pulses
3 are the major cause of Al pulses in soil solution (Johnson, 1992). The connection between Al
4 mobilization and forest decline is not clear. The decline in red spruce certainly has been more
5 severe in the Northeast than in the Southeast, yet all evidence indicates that Al mobilization is
6 most pronounced in the southern Appalachians. However, at the Whiteface Mountain site
7 selected for study because it was in a state of decline, soil solution levels were lower than in the
8 Smokies, which are not in a visibly obvious state of decline (there was no dieback other than the
9 fir killed by the balsam wooly adelgid, no needle yellowing). Thus, Al mobilization constitutes a
10 situation worthy of further study (Johnson, 1992).
11 The simple calculations shown above give some idea of the importance of paniculate
12 deposition in these forest ecosystems, but they cannot account for the numerous potential
13 feedbacks between vegetation and soils nor for the dynamics through time that can influence the
14 ultimate response. One way to examine some of these interactions and dynamics is to use
15 simulation modeling. The nutrient cycling model (NuCM) has been developed specifically for
16 this purpose and has been used to explore the effects of atmospheric deposition, fertilization, and
17 harvesting on some of the IPS sites (Johnson et al., 1993). The NuCM model is a stand-level
18 model that incorporates all major nutrient cycling processes (uptake, translocation, leaching,
19 weathering, organic matter decay, and accumulation).
20 Johnson et al. (1999) used the NuCM model to simulate the effects of reduced S, N, and
21 base cation (CB) deposition on nutrient pools, fluxes, soil, and soil solution chemistry in two
22 contrasting southern Appalachian forest ecosystems: the red spruce and Coweeta hardwood
23 sites from the IPS project. The scenarios chosen for these simulations included "no change;"
24 50% N and S deposition; 50% CB deposition; and 50% N, S, and CB deposition (50% N, S, CB).
25 The NuCM simulations suggested that, for the extremely acid red spruce site, S and N deposition
26 is the major factor affecting soil solution Al concentrations and CB deposition is the major factor
27 affecting soil solution CB concentrations. The effects of S and N deposition were largely through
28 changes in soil solution SO4"2 and NO3" and, consequently, mineral acid anion (MAA)
29 concentrations rather than through changes in soils. This is illustrated in Figures 4-19a,b and
30 4-20a,b, which show simulated soil solution mineral acid anions, base cations, Al, and soil base
31 saturation in B horizon from in the red spruce site. The 50% S and N scenario caused reductions
April 2002 4-121 DRAFT-DO NOT QUOTE OR CITE
-------
to
o
o
to
to
to
fe
H
6
o
o
H
O
O
H
W
O
o
O
CD
*
O
**—
o
.1
"CD
o
o
500 —
400 —
300 —
200 —
c
O 100 —
0-
Red Spruce
Mineral Acid Anions
— No Change
— 50% N,S
— 50% BC
\
2
\
4
\
6
\
8
10
12 14
Year
16 18
20
22 24
Figure 4-19a. Simulated soil solution mineral acid anions and base cations in the red spruce site with no change, 50% N
and S deposition, and 50% base cation deposition. Redrawn from Johnson et al. (1999).
O
-------
to
o
o
to
-^
to
fe
H
6
o
o
H
c>
o
H
W
O
Red Spruce
200
Base Cations
o
"O 160 -
CD
E?
CD
JZ
O
H—
O
c
.g
"cp
"c
0)
o
c
O
O
120
80
40 -
0
— No Change
— 50% N,S
~ 50% BC
\
2
8
10
12
Year
\
14
\
16
18 20 22 24
Figure 4-19b. Simulated soil solution mineral acid anions and base cations in the red spruce site with no change, 50% N
and S deposition, and 50% base cation deposition. Redrawn from Johnson et al. (1999).
O
-------
to
o
o
to
-^
to
fe
H
6
o
o
H
c>
o
H
W
O
_p
O
CD
O
*+-
O
.0
"-*— •
CD
-*-»
c
O
c
o
O
Red Spruce
500 —
400'
300 —
200.
100 —
Al
^ ; i;.
— No Change
— 50% N,S
-- 50% BC
\
2
\
4
\
6
\
8
10
i i
12 14
Year
16 18 20 22 24
Figure 4-20a. Simulated soil solution Al and soil base saturation in the red spruce site with no change, 50% N and S
deposition, and 50% base cation deposition. Redrawn from Johnson et al. (1999).
O
-------
to
o
o
to
Red Spruce
-^
to
fe
H
6
o
o
H
c>
o
H
W
O
o
E
03
-------
1 in soil solution SO4"2, NO3" and, therefore, MAA concentrations, as expected. This, in turn,
2 caused short-term reductions in base cation concentrations. However, by the end of the 24-year
3 simulation, base cations in the 50% S, N scenario were nearly as high as in the no change
4 scenario because base saturation had increased and the proportion of cations as Al decreased.
5 The 50% CB scenario had virtually no effect on soil solution SO4"2, NO3" and, therefore, MAA
6 concentrations, as expected, but did cause a long-term reduction in base cation concentrations.
7 This was caused by a long-term reduction in base saturation (Figure 4-20a,b). Thus, the effects
8 of CB deposition were solely through changes in soils rather than through changes in soil solution
9 MAA, as postulated by Driscoll et al. (1989). In the less acidic Coweeta soil, base saturation was
10 high and little affected by the scenario (cited above); Al was unimportant; and S and
11 N deposition had a much greater effect than CB deposition in all respects (Figure 4-21a,b).
12 In summary, Johnson et al. (1999) found that the results of the red spruce simulations
13 support the hypothesis of Driscoll et al. (1989) in part: CB deposition can have a major effect on
14 CB leaching through time in an extremely acid system. This effect occurred through changes in
15 the soil exchanger and not through changes in soil solution MAA concentration. On the other
16 hand, S and N deposition had a major effect on Al leaching at the Noland Divide site. This
17 occurred primarily because of changes in soil solution MAA concentration. At the less acidic
18 Coweeta site, CB deposition had a minor effect on soils and soil solutions; whereas S and N
19 deposition had delayed but major effects on CB leaching because of changes in SO4"2 and MAA
20 concentrations.
21
22 Effects of Trace Elements
23 Trace metals are natural elements that are ubiquitous in small (trace) amounts in soils,
24 ground water, and vegetation. Many are essential elements required for growth by plants and
25 animals as micronutrients. Naturally occurring surface mineralizations can produce metal
26 concentrations in soils and vegetation that are as high, or higher, than those in the air and
27 deposited near man-made sources (Freedman and Hutchinson, 1981). The occurrence and
28 concentration of trace metals in any ecosystem component depend on the sources of the metal
29 (i.e., via the soil or as a particulate). Even when air pollution is the primary source, continued
30 deposition can result in the accumulation of trace metals in the soil (Martin and Coughtrey,
April 2002 4-126 DRAFT-DO NOT QUOTE OR CITE
-------
>
to
o
o
to
-^
to
fe
H
6
o
o
H
c>
o
H
W
O
o
C
_o
'-I-J
(C
"c
CD
O
c
o
O
Coweeta
0)
05
OS
JZ
O
100
_p
O
E 80
60 -
40
20
Mineral Acid Anions
— No Change
- 50% N,S
~50%BC
I
2
i
4
i
6
i
8
i
10
i
12
i
14
i
16
i
18
20
i
22
2
Year
Figure 4-21a. Simulated soil solution mineral acid anions and base cations in the Coweeta site with no change, 50% N and S
deposition, and 50% base cation deposition. Redrawn from Johnson et al. (1999).
O
-------
to
o
o
to
to
oo
fe
H
6
o
o
H
O
O
H
W
O
Coweeta
100
p
o
E
^ 80
-------
1 1981). Many metals are deposited into soils by chemical processes and are not available to
2 plants (Saunders and Godzik, 1986).
3 When aerial deposition is the primary source of metal particles, both the chemical form and
4 particle size deposited determine the heavy metal concentration in the various ecosystem
5 components (Martin and Coughtrey, 1981). Human activities introduce heavy metals into the
6 atmosphere and have resulted in the deposition of antimony, cadmium, chromium, copper, lead,
7 molybdenum, nickel, silver, tin, vanadium, and zinc (Smith, 1990c). Extensive evidence
8 indicates that heavy metals deposited from the atmosphere to forests accumulate either in the
9 richly organic forest floor or in the soil layers immediately below, areas where the activity in
10 roots and soil is greatest. The greater the depth of soil, the lower the metal concentration. The
11 accumulation of metal in the soil layers where the biological activity is greatest, therefore, has the
12 potential to be toxic to roots and soil organisms and to interfere with nutrient cycling (Smith,
13 1990e). The shallow rooted plant species are those most likely to take up metals from the soil
14 (Martin and Coughtrey, 1981). Though all metals can be directly toxic at high levels, only
15 toxicity from copper, nickel, and zinc have been documented frequently. Toxicity of cadmium,
16 cobalt, and lead has been seen only under unusual conditions (Smith, 1990c). Exposures at lower
17 concentrations have the potential, over the long-term, for interfering with the nutrient-cycling
18 processes when they affect mycorrhizal function.
19 Biological accumulation of metals through the plant-herbivore and litter-detrivore chains
20 can occur. A study of the accumulation of cadmium, lead, and zinc concentrations in
21 earthworms suggested that cadmium and zinc were concentrated, but not lead. Studies indicate
22 that heavy metal deposition onto the soil, via food chain accumulation, can cause excessive
23 levels and toxic effects in certain animals. Cadmium appears to be relatively mobile within
24 terrestrial food chains; however, the subsequent mobility of any metal after it is ingested by a
25 herbivorous animal depends on the site of accumulation within body tissues. Although food
26 chain accumulation may not in itself cause death, it can reduce the breeding potential in a
27 population (Martin and Coughtrey, 1981).
28 In actual case studies, it was observed that the deposition of copper and zinc particles
29 around a brassworks resulted in an accumulation of incompletely decomposed litter. In one
30 study, litter accumulation was reported up to 7.4 km from the stack of a primary smelter in
31 southeastern Missouri. Similar results were reported around a metal smelter at Avonmouth,
April 2002 4-129 DRAFT-DO NOT QUOTE OR CITE
-------
1 England. In the latter case, litter accumulation was associated closely with concentrations
2 specifically of cadmium, as well as with those of lead, copper, and zinc (Martin and Coughtrey,
3 1981). Experimental data (using mesh bags containing litter) supports the hypothesis that
4 reduced decomposition occurs close to heavy metal sources.
5 Accumulations of metals emitted in particulate matter also were reported in soil litter close
6 to a metal smelter at Palmerton, PA, in 1975 and 1978. The continued presence of cadmium,
7 lead, zinc, and copper in the upper soil horizons (layers) were observed 6 years after the smelter
8 terminated operation in 1980. Metal levels were highest near the smelter. The relationship of
9 decreasing amounts of metal in body tissues also held true for amphibians and mammals. Levels
10 of cadmium in kidneys and liver of white-tailed deer (Odocoileus virginaus) were five times
11 higher at Palmerton than in those collected 180 km southwest downwind. The abnormal
12 amounts of metal in the tissues of terrestrial vertebrates and the absence or low abundance of
13 wildlife at Palmerton indicated that ecological processes within 5 km of the zinc smelter
14 continued to be markedly influenced even 6 years after its closing (Storm et al., 1994).
15 Accumulation of heavy metals in litter presents the greatest potential for interference with
16 nutrient cycling. Accumulation of metals in the litter occurs chiefly around brass works and lead
17 and zinc smelters. There is some evidence that invertebrates inhabiting soil litter do accumulate
18 metals. Earthworms from roadsides were shown to contain elevated concentrations of cadmium,
19 nickel, lead, and zinc; however, interference with earthworm activity was not cited (Martin and
20 Coughtrey, 1981). It has been shown, however, that when soils are acidic, earthworm abundance
21 decreases, and bioaccumulation of metals from soil may increase exponentially with decreasing
22 pH (Lokke et al, 1996). Organisms that feed on earthworms living in soils with elevated levels
23 of Cd, Ni, Pb, and Z for extended periods could accumulate lead and zinc to toxic levels (Martin
24 and Coughtrey, 1981). Increased concentrations of heavy metals have been found in a variety of
25 small mammals living in areas with elevated heavy metal concentrations in the soils.
26 Studies by Babich and Stotsky (1978) support the concept that increased accumulation of
27 litter in metal-contaminated areas is due to the effects on the microorganismal populations.
28 Cadmium toxicity to microbial populations was observed to decrease and prolong logarithmic
29 rates of microbial increase, to reduce microbial respiration and fungal spore formation and
30 germination, to inhibit bacterial transformation, and to induce abnormal morphologies. Also, the
31 effects of cadmium, copper, nickel, and zinc on the symbiotic activity of fungi, bacteria, and
April 2002 4-130 DRAFT-DO NOT QUOTE OR CITE
-------
1 actinomycetes were reported by Smith (1991). The formation of mycorrhizae by Glomus
2 mosseae with onions was reduced when zinc, copper, nickel, or cadmium was added to the soil.
3 The relationship of the fungus with white clover, however, was not changed. It was suggested
4 that the effect of heavy metals on vesicular-arbuscular mycorrhizal fungi will vary from host to
5 host (Gildon and Tinker, 1983). Studies with ericoid plants indicated that, in addition to Calluna
6 vulgar is, mycorrhizae also protect Vaccinium macrocarpa and Rhodendron ponticum from heavy
7 metals (Bradley et al., 1981). Heavy metals tend to accumulate in the roots, and shoot toxicity is
8 prevented.
9 The effects of sulfur deposition on litter decomposition in the vicinity of smelters also must
10 be considered. Metal smelters emit SO2 as well as heavy metals. Altered litter decomposition
11 rates have been well documented near SO2 sources (Prescott and Parkinson, 1985). The presence
12 of sulfur in litter has been associated with reduced microbial activity (Bewley and Parkinson,
13 1984). Additionally, the effects on symbiotic activity of fungi, bacteria, and actinomycetes were
14 reported by Smith (1990d).
15 The potential pathways of accumulation of trace metals in terrestrial ecosystems, as well as
16 the possible consequences of trace metal deposition on ecosystem functions, is summarized in
17 Figure 4-22. The generalized trophic levels found in an ecosystem and the various physiological
18 and biological processes that could be affected by trace metals are shown in the figure.
19 Reduction in physiological processes can affect productivity, fecundity, and mortality (Martin
20 and Coughtrey, 1981). Therefore, any effects on structure and function of an ecosystem are
21 likely to occur through the soil and litter (Tyler, 1972).
22 Certain species of plants are tolerant of metal contaminated soils (e.g., soils from mining
23 activities) (Antonovics et al., 1971). Certain species of plants also have been used as
24 bioindicators of metals (e.g., Astragalus is an accumulator of selenium). The sources of both
25 macroelements and trace metals in the soil of the Botanical Garden of the town of Wroclow,
26 Poland, were determined by measuring the concentrations of the metals in Rhododendron
27 catawbiense, Ilex aquifolium, and Mahonia aquifolium growing in the garden and comparing the
28 results with the same plant species growing in two other botanical gardens in nonpolluted areas.
29 Air pollution deposition was determined as the source of metals in plants rather than the soil
30 (Samecka-Cymerman and Kempers, 1999).
31
April 2002 4-131 DRAFT-DO NOT QUOTE OR CITE
-------
1 . Wet/dry deposition
3. Litterfall, resuspension,
deposition, leaching,
stem flow
Biologically
Unavailable
IX.
Soil Organic
X.
Primary
Minerals
11. Mineralization
^
•w
12. Weathering
Atmosphere
1
II.
Plant Surface
Phyllosphere
1
Biologically .
Available /
IV
Upper Soil
x 1 7. Leac
VII.
Lower Soil
j/ ej^' [ 9. Retranslocatio
,-*~^~ III. ?-——?
; 1r Above- -
— > o r ou no — r
2. Foliar uptake ^:~^ Storage, -3^r;
^j~ MGtQbolism
''^ ^-r"~--^r' A "i 1 "' ^-^^ ^J
8. Volatile loss f ~ _J* *\_ ~?-—^~
f ' /* ^^_^/ II iM. Translocation
/^^^ * ^\\^~~~~^^' Root
'^•~'^i/^\ \L \C\turnover
5. Mass flow, V /U- ^ ^
"v^diffusion ^~
^^ V. VI.
y ,f Rhizoplane 6 up0t°ke Metabolism
5. Mass flow,
diffusion
X 7. Leaching
VIII.
Groundwater
Figure 4-22. Relationship of plant nutrients and trace metals with vegetation.
Compartments (roman numerals) represent potential storage sites;
whereas arrows (arabic numerals) represent potential transfer routes.
1 The effects of lead in ecosystems are discussed in the Air Quality Criteria for Lead
2 (U.S. Environmental Protection Agency, 1986). Studies have shown that there is cause for
3 concern in three areas in which ecosystems may be extremely sensitive to lead: (1) delay of
4 decomposition because the activity of some decomposer microorganisms and invertebrates is
5 inhibited by lead, (2) subtle shifts toward plant populations tolerant of lead, and (3) lead in the
6 soil and on the surfaces of vegetation which may circumvent the processes of biopurification.
7 The problems cited above arise because lead is deposited on the surface of vegetation,
8 accumulates in the soil, and is not removed by the surface and ground water of the ecosystem
9 (U.S. Environmental Protection Agency, 1986).
April 2002 4-132 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.2.3 Urban Ecosystems
2 Humans dominate Earth's ecosystems. Their influence on the environment has been
3 pervasive for thousands of years. Evidence has been accumulating from anthropological and
4 archeological research that human influence has been pervasive for thousands of years (Grimm
5 et al., 2000). Major human impacts on the environment probably began as early as 12,000 to
6 15,000 years ago. The Earth abounds with both subtle and pronounced evidence of the influence
7 of humans on natural ecosystems. Clearly, human actions have continued to dramatically alter
8 the functioning of ecosystems of which they are a part, and, it is equally clear that humans now
9 influence virtually all ecosystems. Nowhere has human action been more intense than in cities,
10 suburbs, exurbs and in the supporting hinterlands (Grimm et al., 2000). This fact has lead to
11 much recent interest in the study of ecological systems.
12 Vitousek et al. (1997) point out that understanding a human-dominated planet requires that
13 the human dimensions of global change—the social, cultural, and other drivers of human
14 actions—need to be included within ecological analyses. Therefore, humans must be integrated
15 into models for a complete understanding of extant ecological systems. Development of more
16 realistic models for ecological systems will lead to greater success in finding solutions to
17 environmental problems.
18 In the past, ecological plant or animal studies conducted in urban settings used traditional
19 ecological approaches and considered humans as agents of disturbance. Although the term urban
20 ecosystem has been used to describe human-dominated ecosystems, it does not adequately take
21 into account the developmental history, sphere of influence, and potential impacts required in
22 order to understand the true nature of an urban ecosystem (Mclntyre, et al., 2000). Urban
23 ecology, because urbanization is both an ecological and a social phenomenon, implicitly
24 recognizes the role humans play in developing unique systems. Therefore, if urban ecology is to
25 be a truly interdisciplinary field, it is crucial that it integrate both social and natural sciences into
26 the study of urban ecosystems (Mclntyre, et al., 2000).
27 Although the study of ecological phenomena in urban environments is not a new area of
28 science, the concept of the city as an ecosystem is relatively new for the field of ecology (Grimm
29 et al., 2000). There is a wealth of information on the terrestrial components of urban ecological
30 systems. However, much of it is organized from the perspective of ecology in cities while the
31 more comprehensive perspective identified as ecology of cities is needed (Pickett et al., 2001).
April 2002 4-133 DRAFT-DO NOT QUOTE OR CITE
-------
1 The basic questions addressed by the literature of ecology in cities are how do ecological patterns
2 and processes differ in cities as compared with other environments? What is the effect of the city
3 (i.e., a concentration of human population and activities) on the ecology of organisms inside and
4 outside of its boundary and influence? The concept of ecology of cities has to do with how
5 aggregated parts make up the whole, i.e., how cities process energy or matter relative to their
6 surroundings (Grimm et al., 2000). The latter concept includes primary production, species
7 richness, biogeophysical budgets, ecosystem patterns and processes, and an open definition of
8 urban ecosystems that incorporates the exchanges of materials and influence between cities and
9 surrounding landscapes (Pickett et al., 2001). If ecosystems are to be understood, there is a need
10 for a new integrative ecology that explicitly incorporates human decisions, culture, institutions,
11 and economic systems (Grimm et al., 2000). This fact makes an ecological approach to land use
12 planning not only necessary but essential to maintain long-term sustainability of ecosystem
13 benefits, services, and resources (Zipperer et al., 2000). The ecological and social effects of
14 "edge city" need to be studied as they may be greater than the previous patterns of
15 suburbanization. The classical ecosystem approach and a patch dynamic approach are needed to
16 understand and manage the dynamics of urban and urbanizing ecosystems (Zipperer et al., 2000).
17 There has been little work on the rates of atmospheric deposition to urban ecosystems
18 despite the large body of knowledge on concentrations and chemical reactions of air pollutants in
19 cities. A search of the abundant literature produced no references that dealt with the effects of
20 PM deposition. Lovett et al. (2000), however, reported that urban ecosystems are likely to be
21 subjected to large rates of deposition of anthropogenic pollutants. Decades of research on urban
22 air quality indicate that cities are often sources of nitrogen oxides, sulfur oxides, and dust, among
23 many other pollutants. Some of these air pollutants are major plant nutrients (e.g., nitrogen) and
24 may be affecting nutrient cycles in plant-dominated areas in and around cities. Studying the
25 deposition rates of atmospheric pollutants in urban areas can provide a quantitative estimate of
26 the amounts of gaseous and particulate air pollutants that are removed by urban vegetation.
27 To determine the patterns of atmospheric deposition and throughfall in the vicinity of a
28 large city, Lovett et al. (2000) measure bulk deposition, oak forest throughfall, and particulate
29 dust at sites along a transect within and to the north of New York City. The gases and particles
30 in urban air can increase atmospheric deposition within and downwind of the city. They
31 observed that concentrations and fluxes of NO3", NH4+, Ca+2, Mg+2, SO4"2, and Cl" in throughfall
April 2002 4-134 DRAFT-DO NOT QUOTE OR CITE
-------
1 all declined significantly with distance from the city, while hydrogen ion concentration and flux
2 increased significantly with distance from the city. Most of the change in concentrations and
3 fluxes occurred within 45 km of the city. Also, it was observed that throughfall nitrogen was
4 twice as high in the urban areas when compared with the suburban and rural sites. Most of the
5 dry deposition of nitrate was from gaseous nitrogen oxides. As mentioned above, the effects of
6 the atmospheric deposition of the particulate pollutants was not mentioned.
7 McDonnell et al. (1997) in a 10-year study of ecosystem processes along an urban-rural
8 gradient included plant litter dynamics and nitrogen cycling of two key components of a forest
9 ecosystem: litter decomposition and heavy metal levels in soil and foliar litter. Foliar litter
10 decomposition integrates many features of the abiotic and biotic environment. It is an important
11 site of heavy metal incorporation into ecosystems and provides a both a habitat and a resource for
12 fungi, bacteria, and invertebrates. Litter decomposition integrates the effects of resource quality,
13 environmental factors, and activities of decomposer organisms on nutrient cycling and serves as
14 an easily measured indicator of the effect of urbanization on an important ecosystem function.
15 McDonnell et al. (1997) noted that levels of heavy metals in the foliar litter in urban forest soils
16 were higher than in rural. The levels in urban forest stands approached or exceeded the levels
17 reported to affect soil invertebrates, macrofungi, and soil microbial processes. The urban forests
18 exhibited reduced fungal biomass and microarthropod densities when compared to rural stands.
19 These results supported the concept that urban forests have depauperate communities because of
20 anthropogenic stress resulting from poor air quality due to high levels of SO2, sulfate, ozone and
21 nitrogen; elevated levels of soil and forest floor heavy metals; and low water availability such as
22 those caused by hydrophobic soils (McDonnell et al.,1997). Thus, forests at the urban end of the
23 gradient exhibited reduced fungal and microarthropod populations and poorer leaf quality than
24 the more rural forests. The potential effect of these conditions on the ecosystem processes of
25 decomposition and nitrogen cycling in urban forests appeared to be ameliorated by two other
26 anthropogenic factors: increased average temperatures caused by the heat island effect and the
27 introduction and successful colonization of earthworms in the urban forests (McDonnell
28 et al.,1997).
29 McDonnell et al. (1997) observed that the changes in forest nitrogen dynamics were related
30 to increased anthropogenic nitrogen deposition in an urban environment. The studies of Aber
31 et al. (1989) in the northeastern United States on forest nitrogen dynamics demonstrated that
April 2002 4-135 DRAFT-DO NOT QUOTE OR CITE
-------
1 elevated nitrogen deposition over many years results in increased nitrification and the
2 mineralization of more nitrogen than can be taken up by plants and microorganisms.
3 Nitrification can precipitate decreases in fine root biomass and increases in nitrate leaching
4 below the root zone. These effects of nitrogen deposition were not related to inputs from a
5 specific source such as PM.
6 There have also been studies of heavy metal deposition in or near cities. However, the
7 studies do not cite the effects of metals in the soil. Pouyat and McDonnell (1991) discuss heavy
8 metal accumulations in forest soils along an urban-rural gradient in southeastern New York.
9 Variations in the amounts of Zn, Cu, Ni, and Cd appeared indicative of a pattern of atmospheric
10 deposition near point sources (Section 4.3.2.6). The concentrations of heavy metals in forest
11 floor and soils corresponded closely with the urban-rural land use gradient. Again, as in the
12 study by Lovett et al. (2000), the pollutants were highest near the urban end of the gradient and
13 declined toward rural sites with Pb, Ni and Cu highest near the urban end.
14 The air quality of the region around East St. Louis has been a source of concern due the
15 industries in the area (Kaminski and Landsberger, 2000a). Industries include ferrous and
16 nonferrous metal smelters (Pb, Zn, Cu, and Al), coal-fired power plants, producers of organic and
17 inorganic chemicals, municipal waste incinerators, and petroleum refineries. The city also is
18 located in the path of diverse plumes from refineries to the north, coal-fired power plants to the
19 west, and nonferrous smelters to the south. Due to years of exposure to the industrial emissions,
20 concerns have arisen with the community about the environmental impact. Concentrations of
21 heavy metals and metalloids (As, Cd, Cu, Hg, Pb, Sb, Zn) in the soil provided a basis for analysis
22 (Kaminski and Landsberger, 2000b). The dual aims of these studies was (1) to make an accurate
23 technical assessment of the extent of the pollutants on the soil surface as well as the extent of the
24 depth to determine possibilities of remediation and (2) to determine the leaching dynamics of
25 heavy metals to determine possible effects on biota uptake or groundwater contamination. The
26 effects on biota are not mentioned; however, the soils in the area acted as a sink and there was
27 little groundwater mobility (Kaminski and Landsberger, 2000b). The possible effects of heavy
28 metals in soils is discussed in the previous section (4.3.2.6) on trace metals.
29 The above assessment of new information leads to the clear conclusion that atmospheric
30 PM at levels currently found in the United States have the potential to alter ecosystem structure
April 2002 4-136 DRAFT-DO NOT QUOTE OR CITE
-------
1 and function in ways that may reduce their ability to meet societal needs. The possible direct
2 effects of airborne PM on individual plants were discussed in Section 4.2.2.1.
3
4 4.2.4 Ecosystem Goods and Services and Their Economic Valuation
5 Human existence on this planet depends on ecosystems and the services and products they
6 provide. The essential services and products provided by the planet's collective biodiversity (the
7 earth's flora, fauna, and microorganisms) are clean air, clean water, clean soil, and clean energy
8 (Table 4-17). Today, governments around the world tend to pursue a "bottom line" that is driven
9 by an economy that is disconnected from the natural world and is fundamentally destructive of
10 local ecosystems (Suzuki, 1997). For this reason, human society needs to be reconnected to the
11 biologically diverse ecosystems so that they realize that they are a part of the natural world
12 (Suzuki, 1997). There is a need to understand that biodiversity encompasses all levels of
13 biological organization, including individuals, populations, species, and ecosystems (Wilson,
14 1997). Human-induced changes in biotic diversity and alterations in the structure and
15 functioning of ecosystems are the two most dramatic ecological trends in the past century
16 (Vitousek et al., 1997). Ecosystem processes such productivity, nitrogen mineralization rate, and
17 nitrate leaching respond directly to human modification of ecosystems and to changes in
18 atmospheric composition and climate (Chapin et al., 1997). Habitat conversion, changes in land
19 use, and the introduction of exotic species result in changes in biota, reduced genetic/species
20 diversity, and leads to a homogenization of the global biota. These biotic changes will influence
21 ecosystems processes sufficiently to alter the future state of the world's ecosystems and the
22 services they provide (Chapin et al., 1997).
23 Though Homo sapiens is only one of perhaps 5-30 million animal species on earth, it
24 controls a disproportionate share of the planet's resources. Humans are co-opting approximately
25 40% of the present net primary production (NPP) of organic material each year. NPP is the
26 amount of energy remaining after subtracting the respiration of primary producers (mostly plants)
27 from the total amount of energy (mostly solar) that is fixed biologically and provides the basis for
28 the maintenance, growth and reproduction of all consumer and decomposers. It is the total food
29 resource of earth (Vitousek et al., 1986).
30 The number, biodiversity, structure, and functions of ecosystem populations, provide
31 ecosystem products (goods) and services (Figure 4-23). For any given population, the number of
April 2002 4-137 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-17. PRIMARY GOODS AND SERVICES PROVIDED BY ECOSYSTEMS
Ecosystem
Goods
Services
Agroecosvstems
Coastal ecosystems
Forest ecosystems
Freshwater
Grassland
ecosystems
Food crops
Fiber corps
Crop genetic resources
Fish and shellfish
Fishmeal (animal feed)
Seaweeds (for food and
industrial use)
Salt
Genetic resources
• Timber
• Fuelwood
• Drinking and irrigation water
• Fodder
• Nontimber products (vines,
bamboos, leaves, etc.)
• Food (honey, mushrooms fruit,
and other edible plants; game)
• Genetic resources
Drinking and irrigation water
Fish
Hydroelectricity
Genetic resources
Livestock (food, game, hides,
and fiber)
Drinking and irrigation water
Genetic resources
• Maintain limited watershed functions (infiltration,
flow control, and partial soil protection)
Provide habitat for birds, pollinators, and soil
organisms important to agriculture
• Sequester atmospheric carbon
Provide employment
• Moderate storm impacts (mangroves, barrier
islands)
• Provide wildlife (marine and terrestrial) habitat
and breeding areas/hatcheries/nurseries
• Maintain biodiversity
Dilute and treat wastes
Provide harbors and transportation routes
Provide human and wildlife habitat
Provide employment
Contribute aesthetic beauty and provide recreation
Remove air pollutants, emit oxygen
• Cycle nutrients
• Maintain array of watershed functions
(infiltration, purification, flow control, soil
stabilization)
• Maintain biodiversity
• Sequester atmospheric carbon
• Moderate weather extremes and impacts
• Generate soil
Provide employment
Provide human and wildlife habitat
Contribute aesthetic beauty and provide recreation
• Buffer water flow (control timing and volume)
Dilute and carry away wastes
• Cycle nutrients
• Maintain biodiversity
Provide aquatic habitat
Provide transportation corridor
Provide employment
Contribute aethetic beauty and provide recreation
• Maintain array of watershed functions
(infiltration, purification, flow control, and soil
stabilization)
• Cycle nutrients
Remove air pollutants and emit oxygen
• Maintain biodiversity
• Generate soil
• Sequester atmospheric carbon
Provide human and wildlife habitat
Provide employment
Contribute aesthetic beauty and provide recreation
Source: World Resources (2000-2001).
April 2002
4-138
DRAFT-DO NOT QUOTE OR CITE
-------
Water use and
nutrient loss
Hydrologic
CO2 and
temperatun
changes
Land Precipitation/erosion
transfer- and temp- / changes
mation erature / jn water flow
and temperature1
Food supply
and demand
Water availability
N, CH4,N20
emissions
Climate change
Loss of crop
genetic diversity
Habitat loss
Change in
transpiration
albedo
Biodiversity loss
Freshwater
supply and demand
Habitat
change
Forest product
supply and demand
Loss and
fragmentation,
of nabitat
Reduced resilience
to change
Figure 4-23. Linkages among various ecosystem goods and services (food, water,
biodiversity, forest products) and other driving forces (climate change)
(modified from Ayensu et al., 1999).
1 individuals, the genetic variation between individuals, and the area occupied affects ecosystem
2 functioning and the delivery of ecosystem services and other benefits provided by that population
3 (Hughes, et al., 1997). Loss or altering of population diversity means loss of the benefits
4 described in Table 4-17 and, in particular, with time, the loss of the life-support systems on
5 which humanity relies (Hughes et al., 1997).
6 Attempts have been made to calculate the value of biodiversity and the world's ecosystem
7 services and natural capital (Pimentel et al., 1997; Costanza et al., 1997). Pimentel et al. (1997)
8 estimated economic and environmental benefits for services contributed from all biota in the
9 United States, including their genes, at $319 billion per year. Costanza et al. (1997) have
10 estimated the total value of ecosystem services by biome for the entire bioshere and concluded
11 that ecosystems provide at least $33 trillion worth of services annually. Approximately 63% of
April 2002
4-139
DRAFT-DO NOT QUOTE OR CITE
-------
1 the estimated value is contributed by marine ecosystems ($20.9 trillion per year), most of which
2 comes from coastal ecosystems ($10.6 trillion per year). About 38% of the estimated value
3 comes from terrestrial ecosystems, mainly from forests ($4.7 trillion per year) and wetlands
4 ($4.9 trillion per year). Costanza et al. (1997) state that it may never be possible to make a
5 precise estimate of the services provided by ecosystems; however, their estimates indicate the
6 relative importance of ecosystem services, not their true value considering that the loss of
7 ecosystem services can affect human existence.
8 Heal (2000), however, feels that attempts to value ecosystems and their services are
9 misplaced: "Economics cannot estimate the importance of natural environments to society: only
10 biology can do that" (Heal, 2000). The role of economics is to help design institutions that will
11 provide incentives to the public and policy makers for the conservation of important natural
12 systems and for mediating human impacts on the biologically diverse ecosystems and the
13 biosphere so that they are sustainable. The approach of Harwell et al. (1999) also deals with the
14 need to understand human effects on ecosystems so that ecosystem management can define what
15 ecological conditions are desired. Further, they state that the establishment of ecological goals
16 involves a close linkage between scientists and decision makers in which science informs
17 decision makers and the public by characterizing the ecological conditions that are achievable
18 under particular management regimes. Decision makers then can make choices that reflect
19 societal values including issues of economics, politics, and culture. For management to achieve
20 their goals, the general public, scientific community, resource managers, and decision makers
21 need to be routinely apprised of the condition or integrity of ecosystems so that ecological goals
22 may be established (Harwell et al., 1999).
23 Though usually considered as toxic pollutants locally (Section 4.3.2.3), secondary organics
24 as PM can become airborne and distributed over a wide area and affect ecosystems remote from
25 the source. Some of the chemical compounds may reach toxic levels in the food chains of human
26 and other animals. However, other compounds tend either to decrease or maintain the same level
27 of toxicity. The effects of toxicity on the animal population can alter the functioning of the
28 ecosystem. The major impacts of airborne PM on ecosystems, however, are through the indirect
29 effects on plant populations that occur through the soil and affect the cycling of nutrients
30 necessary for plant growth, vigor, and maintenance of biodiversity as discussed in Section
31 4.2.2.2. By altering the cycling of nitrogen, nitrogen deposition changes the biodiversity of
April 2002 4-140 DRAFT-DO NOT QUOTE OR CITE
-------
1 ecosystems and their functioning and, by altering the vigor of forest tree stands, alters forest
2 succession. Additionally, nitrogen deposition in combination with the deposition of sulfur in the
3 form of acid rain alters the biogeochemical cycling of soil mineral nutrients and changes the
4 biodiversity and functioning of forest ecosystems. The changes in the ability of forest vegetation
5 and soil microorganisms to utilize nutrients results in the leaching of nitrates and other minerals
6 from the soils. The nitrate and mineral runoff affects streams and coastal and aquatic ecosystems
7 and, thus, influences the services important to human life provided by these ecosystems as well
8 (Table 4-17).
9
10
11 4.3 EFFECTS ON VISIBILITY
12 4.3.1 Introduction
13 Visibility may be thought of as the degree to which the atmosphere is transparent to visible
14 light (National Research Council, 1993). The beauty of scenic vistas in many parts of the U.S. is
15 often diminished by haze that reduces contrast, washes out colors, and renders distant landscape
16 features indistinct or invisible. This degradation of visibility is due primarily to the scattering
17 and absorption of light by fine particles suspended in the atmosphere. One quantitative measure
18 of visibility, used traditionally by meteorologists, is the visual range, defined as the farthest
19 distance at which a large black object can be distinguished against the horizon sky (U.S.
20 Environmental Protection Agency, 1979).
21 In August 1977, Congress amended the Clean Air Act (CAA) to establish as a national goal
22 "the prevention of any future and remedying of any existing impairment of visibility in
23 mandatory Class I Federal areas (many national parks and wilderness areas), which impairment
24 results from manmade air pollution" (Title I Part C Section 169A, U.S. Code [1990]). The 1977
25 Amendments also included provisions requiring applicants for new major source permits to
26 assess the potential for their projects to cause adverse impacts on the air quality-related values,
27 including visibility, in nearby Class I areas. In 1980, the EPA established regulatory
28 requirements under Section 169A to address Class I protection from "reasonably attributable"
29 visibility impairment, i.e., visibility impairment attributable to a single source or small group of
30 sources.
April 2002 4-141 DRAFT-DO NOT QUOTE OR CITE
-------
1 The CAA, as amended in 1990 (section 169B), required the U.S. Environmental Protection
2 Agency to conduct research on regional visibility impairment and to establish the Grand Canyon
3 Visibility Transport Commission (GCVTC). The GCVTC was charged with assessing and
4 providing recommendations to help preserve clear days and improve visibility in the 16 national
5 parks and wilderness areas located on the Colorado Plateau. The GCVTC was mandated to
6 provide recommendations to the U.S. Environmental Protection Agency for the reduction of
7 visibility impairment due to regional haze, described as any perceivable change in visibility (light
8 extinction, visual range, contrast, or coloration) from that which would have existed under
9 natural conditions that is caused predominantly by a combination of many anthropogenic sources
10 over a wide geographical area (U.S. Environmental Protection Agency, 1999a). In July 1999, the
11 U.S. Environmental Protection Agency published the Regional Haze Rule (Federal Register,
12 1999). The regulation established a program for the improvement and protection of visibility in
13 the 156 protected Class I parks and wilderness areas, including the establishment of baseline and
14 current visibility conditions and the tracking of changes in visibility conditions over time.
15 Implementation of the regional haze regulations is supported by the U.S. Environmental
16 Protection Agency's PM25 monitoring network and an expanded Interagency Monitoring of
17 Protected Visual Environments (IMPROVE) network. A discussion on the PM25 monitoring
18 network and the IMPROVE network appears elsewhere in this section (National Park Service,
19 1998; Evans and Pitchford, 1991; U.S. Environmental Protection Agency, 2000b; U.S.
20 Environmental Protection Agency, 2001).
21 The objective of the visibility discussion in this section is to provide a brief description of
22 the fundamentals of atmospheric visibility and to summarize the linkage between particulate
23 matter and visibility. Visibility is an effect of air quality and, unlike the particulate matter
24 concentration, is not a property of an element of volume in the atmosphere. Visibility can be
25 quantified only for a sight path and depends on the illumination of the atmosphere and the
26 direction of view. However, the concentration of particles in the atmosphere plays a key role in
27 determining visibility. Therefore, visibility impairment may be controlled by control of particle
28 concentrations. The relationships between particles, other factors, and visibility impairment are
29 described in this section. For a more detailed discussion on visibility, the reader is referred to the
30 1996 Air Quality Criteria for Particulate Matter (PM AQCD) (U.S. Environmental Protection
31 Agency, 1996a), the Recommendations of the Grand Canyon Visibility Transport Commission
April 2002 4-142 DRAFT-DO NOT QUOTE OR CITE
-------
1 (Grand Canyon Visibility Transport Commission, 1996), the National Research Council
2 (National Research Council, 1993), the National Acid Precipitation Assessment Program
3 (Trijonis et al., 1991), Interim Findings on the Status of Visibility Research (U.S. Environmental
4 Protection Agency, 1995a), and reports summarizing visibility science and data from the
5 IMPROVE visibility monitoring network (Malm et al., 2000; Sisler, 1996; Sisler et al., 1993).
6
7 4.3.2 Factors Affecting Atmospheric Visibility
8 The visual perceptions of a distant object is influenced by a large number of factors
9 including human vision, various characteristics of the atmosphere (e.g., atmospheric
10 illumination, path and transmitted radiance, contrast, and optical properties), and atmospheric
11 pollution. Detailed discussion of this full range of topics can be found in the 1996 PM AQCD
12 (U.S. Environmental Protection Agency, 1996a) and other general references (e.g., Malm, 1999).
13 This section focuses only on those topics that have generally been addressed by more recent
14 research, including atmospheric illumination, the optical properties of gases and particles in the
15 atmosphere, and the effects of relative humidity on the optical properties of particles.
16
17 4.3.2.1 Optical Properties of the Atmosphere and Atmospheric Particles
18 Atmospheric particles and gases attenuate image-forming light as it travels from a viewed
19 object to an observer. The fractional attenuation of light per unit distance is known as the light
20 extinction coefficient. The light extinction coefficient, bext, is expressed in units of one over
21 length, for example inverse kilometers (km"1) or inverse megameters (Mm"1). The light
22 extinction coefficient can be expressed as the sum of the light scattering and light absorption
23 coefficients of particles and gases.
24
25 bext = bap + bag + bsg+ bsp (4-6)
26
27 where the subscripts/? and g signify particles and gases, and s and a signify scattering and
28 absorption.
29
30
31
April 2002 4-143 DRAFT-DO NOT QUOTE OR CITE
-------
1 Transmitted Radiance and Path Radiance
2 The appearance of a distant object is determined by light from two sources: the light
3 reflected from the object itself and the light reflected by the intervening atmosphere. Light
4 reflected by the object is attenuated by scattering and absorption as it travels through the
5 atmosphere toward the observer. The portion that reaches the observer is the transmitted
6 radiance. During the daytime, the sight path is illuminated by the direct rays of the sun, diffuse
7 skylight, light that has been reflected from the surface of the Earth, etc. Some of this
8 illumination is scattered toward the observer by the air molecules and paniculate matter in the
9 sight path. The accumulation of the light scattered into the sight path is the path radiance or air
10 light. The path radiance is significantly influenced by the illumination of the sight path.
11 However, not all of the light scattered into the sight path reaches the observer.
12 The transmitted radiance carries the information about the object. The path radiance only
13 carries information about the intervening atmosphere and is often quite featureless. When the
14 transmitted radiance is dominant, visibility is good. Conversely, when the path radiance is
15 dominant, visibility is poor. In a dense fog, the transmitted radiance from nearby objects can be
16 seen, but the transmitted radiance from more distant objects is completely overwhelmed by the
17 path radiance (i.e., the light scattered by the fog). Distant objects are lost in the white (or gray) of
18 the fog (Gazzi et al., 2001).
19 Figure 4-24 illustrates the radiance seen by an observer looking at a hillside or through the
20 aperture of a measurement instrument. The radiance that enters the eye of the observer (or the
21 aperture of a measurement instrument) is known as the apparent radiance (i.e., the sum of the
22 transmitted and path radiance). It is the competition between the transmitted radiance and the
23 path radiance that determines visibility.
24
25 Light Absorption and Scattering by Gases
26 In the ambient atmosphere the only visible light absorbing gas of any consequence is
27 nitrogen dioxide (NO2), which absorbs primarily blue light, thus causing a yellow or brown color
28 if in sufficient concentration across a sight path. Usually the absorption by NO2 is much smaller
29 than the scattering by particles that are typically present in polluted environments, such as urban
30 areas. The most common exception to this situation of relatively small NO2 absorption is in
31 effluent plumes from combustion facilities where the particles are effectively removed but the
April 2002 4-144 DRAFT-DO NOT QUOTE OR CITE
-------
OBSERVER
Figure 4-24. Light reflected from a target toward an observer. The intervening
atmosphere scatters a portion of this light out of the sight path and scatters
light from the sun into the sight path. Some particles and gases also absorb a
portion of the light from the target. The light scattered into the sight path
increases with distance from the target, whereas the light transmitted from
the target decreases with distance from the target. The visual range is the
closest distance between the target and the observer at which the transmitted
light no longer can be distinguished from the light scattered into the sight
path.
Source: Watson and Chow (1994).
1
2
3
4
5
6
7
nitrogen oxide (NO), which can convert rapidly to NO2, is not removed. Except for such
particle- depleted NO plumes, the light absorption coefficient for gases is usually ignored in
determinations of the light extinction coefficient.
Scattering by gases in the atmosphere is described by the Rayleigh scattering theory
(vandeHulst, 1981) and is referred to as Rayleigh scattering. The magnitude of the Rayleigh
scattering depends on the gas density of the atmosphere and varies from about 9 Mm"1
to 11 Mm"1 for most locations of interest, depending primarily on site elevation. To simplify
April 2002
4-145
DRAFT-DO NOT QUOTE OR CITE
-------
1 comparisons of light extinction coefficient values among sites at a variety of elevations, a
2 standard value of 10 Mm"1 is often used for the Rayleigh scattering component (Malm, 2000).
3
4 Light Absorption by Particles
5 Absorption by particles is principally caused by elemental carbon (also referred to as soot
6 or light-absorbing carbon), which usually results from incomplete combustion of fossil fuels.
7 Some minerals in crustal particles also absorb light and can be a significant factor during fugitive
8 dust episodes.
9 Most particle absorption data are determined by measuring light transmission of particles
10 captured on filter media. Absorption estimates made in this way are sensitive to the filter
11 substrate used, the optical configuration of the transmission measurement, particle loading on the
12 filter, and particle scattering albedo, with the result that there are significant uncertainties for
13 measurements of filtered particles (Horvath, 1993). Another approach to estimating aerosol light
14 absorption is by subtracting concurrent light scattering measurements, using a nephelometer,
15 from light extinction measurements, using a transmissometer. Substantial uncertainties in this
16 difference approach result from the assumption that the point measurement of scattering is
17 representative of the scattering over a long path (1 to 10 km), typically required for
18 transmissometers measurements. A recently field-tested prototype photoacoustic spectrometer
19 designed to determine absorption of suspended aerosol and an enclosed-folded path
20 transmissometer offers hope of resolving the problems of the filter-based and difference
21 approaches to the measurement of light absorption by particles (Arnott et al., 1999).
22 The relationship between elemental carbon concentration and particle absorption can be
23 calculated using Mie equations for particles with known size distribution, particle density, index
24 of refraction, shapes, and for various internal mixtures with non-absorbing aerosol materials
25 (Fuller et al., 1999). While such application of this theory can provide a range of absorption
26 efficiencies for various model aerosol distributions, it is rare that sufficiently detailed particle
27 characterization data for ambient aerosols are available. Also, although elemental carbon is the
28 strongest and most common of the absorbing particles, light absorption by elemental carbon
29 particles can be reduced when the particle is covered by other chemical species (Dobbins et al.,
30 1994) or may be enhanced when coated with a non-absorbing refractive material such as
31 ammonium sulfate (Fuller et al., 1999).
April 2002 4-146 DRAFT-DO NOT QUOTE OR CITE
-------
1 More commonly, estimates of elemental carbon absorption efficiency are empirically
2 determined from the ratios of or the slopes of regression analysis fits to absorption coefficient
3 and corresponding elemental carbon concentration measurements. Use of the regression
4 approach permits the inclusion of crustal component concentrations as a second dependent
5 parameter, so that crustal absorption can also be estimated. Uncertainties in the absorption
6 efficiency determined empirically are a combination of the measurement uncertainties for the
7 absorption coefficients, elemental carbon concentrations, and where used, the crustal
8 concentrations. In reviews of estimates of elemental carbon light absorption mass efficiency (i.e.,
9 the absorption coefficient per carbon mass concentration), Horvath (1993) and Liousse et al.
10 (1993) found values ranging from 2 to 17 m2/g. Moosmiiller et al. (1998) showed that by
11 limiting the absorption coefficient estimates to those using photoacoustic methods, the
12 absorption efficiency shows a wavelength dependence, with highest values (17 m2/g) at the
13 shortest wavelength used (A = 0.42 //m) and lowest values (3 m2/g) at the longest wavelengths
14 used (A = 0.8 //m). The center of the visible light wavelength (A = 0.53 //m) yielded elemental
15 carbon absorption efficiencies values of about 10 m2/g, which is a commonly used value for
16 elemental carbon absorption efficiency. Fuller et al. (1999) suggested that isolated spheres of
17 light absorbing carbon have a specific absorption of less than 10 m2/g. Light absorption by
18 carbon particles only will be greater than 10 m2/g if the particles are internally mixed and the
19 occluding particles are sufficiently large. Absorption values for graphitic and amorphous carbon
20 for primary sizes typical of diesel soot are around 5 m2/g.
21
22 Light Scattering by Particles
23 Particle scattering tends to dominate light extinction, except under pristine atmospheric
24 conditions when Rayleigh scattering by gas molecules is the largest contributor. The Mie
25 equation can be used to calculate particle scattering for aerosols of known size distribution,
26 particle density, index of refraction, shape, and for various known internal component mixtures
27 (Fuller et al., 1999). Unlike particle absorption, which is principally associated with elemental
28 carbon, all particles scatter light.
29 Light-scattering by particles has been reported to account for 68 to 86% of the total
30 extinction coefficient in several cities in California (Eldering et al., 1994). When light-scattering
31 increases, visibility is impaired because of a decrease in the transmitted radiance and an increase
April 2002 4-147 DRAFT-DO NOT QUOTE OR CITE
-------
1 in the path radiance. The single most important factor that determines the amount of light
2 scattered by a particle is its size, as shown in Figure 4-25. The maximum single particle
3 scattering efficiency (i.e., scattering per cross-sectional area of the particle) is associated with
4 particles with diameters of about the wavelength of visible light (centered at 0.53 //m).
5 For particles that are small compared to the wavelength of light, the single particle scattering
6 efficiency is low. For particles larger than the wavelength, the single particle scattering
7 efficiency initially decreases with diameter and then fluctuates around a value of two as size
8 increases. However, a larger particle always scatters more light than a smaller particle because
9 particle cross-sectional area increases faster with diameter than the decrease in single particle
10 scattering efficiency at any point on the scattering efficiency curve. The mass scattering
11 efficiency (i.e., the scattering per mass concentration) peaks for particles that are about 0.5 //m to
12 0.8 //m in diameter. Smaller particles are much less efficient at scattering light, while larger
13 particles have mass that increases with particle size faster than the increase in the amount of light
14 they scatter.
15
16
4.5
O 4.0 r"\A Total MIE Scattering Coefficient
CD ^ 35 / \ ForR=1.50
"o o>
t 'O 3.0
CD it
Q. LLJ 2.5
-------
1 Use of the Mie equation to calculate light scattering or the light scattering efficiency of
2 particles in the atmosphere is severely limited by the general lack of sufficiently detailed particle
3 characterization data. At a minimum, size-resolved particle composition data (e.g., aerosol
4 collected on an 8-stage impactor) are needed to permit meaningful Mie scattering calculations.
5 The chemical composition provides clues to the appropriate particle density and index of
6 refraction, while the size distribution is inferred by fitting a distribution function to the
7 concentration for each stage. Assumptions are still necessary to address the particle component
8 mixture characteristics of the aerosol. Resulting scattering calculations can be compared to
9 directly measured particle extinction to assess the reasonableness of the Mie calculations.
10 Reported calculated scattering efficiencies for sulfates range from 1.2 to 5.6 m2/g. Sulfate
11 scattering efficiencies have been reported to increase by a factor of two when the size distribution
12 went from 0.15 to 0.5 //m (McMurry et al., 1996). Calculated scattering efficiencies for carbon
13 particles ranged from 0.9 to 8.1 m2/g. A scattering efficiency of 1.0 and 0.6 m2/g was reported
14 for soil and coarse mass, respectively (U.S. Environmental Protection Agency, 1996a; Sisler and
15 Malm, 2000).
16 Integrating nephelometers directly measure ambient particle scattering. A variety of
17 nephelometer configurations that can include the use of unrestricted or size selective inlets and
18 the control of sample air temperature and relative humidity permit the composite scattering
19 properties of ambient aerosol to be directly observed (Day et al., 1997). When sample-controlled
20 nephelometer data are combined with collocated particle speciation data, composite particle
21 scattering efficiency values for ambient aerosol can be empirically derived (Malm et al., 2000).
22
23 4.3.2.2 Relative Humidity Effects on Particle Size and Light-Scattering Properties
24 The ability of some commonly occurring chemical components of atmospheric aerosol to
25 absorb water from the vapor phase has a significant impact on particle light scattering.
26 Hygroscopic particulate materials, which typically include sulfuric acid, the various ammonium
27 sulfate salts, ammonium nitrate, and sodium chloride, change size by the accumulation and loss
28 of water as they maintain equilibrium with the vapor phase as a function of changes in relative
29 humidity. For some materials (e.g., sulfuric acid), the growth is continuous and reversible over
30 the entire range of relative humidity. For other materials, water absorption begins abruptly for a
31 dry particle at a specific relative humidity known as the deliquescent point (e.g., -80% for
April 2002 4-149 DRAFT-DO NOT QUOTE OR CITE
-------
1 ammonium sulfate) and continues as relative humidity increases. There is a hysteresis effect
2 with these materials in that, once wet, the relative humidity can be reduced below the
3 deliquescent point until crystallization occurs at a substantially lower relative humidity (e.g.,
4 -30% for ammonium sulfate). Figure 4-26 shows the water vapor growth curve for ammonium
5 sulfate.
Q
O
ro
DC
o
O
(D
TO
Q
1.5-
1.0-
B
Hysteresis Loop
for (NH4)2 SO4
M
- 5
>
O
- 4 =
03
ct:
A
- 3
o
2 3
Q)
1 I
30
50
70
90
Relative Humidity (%)
Figure 4-26. Particle growth curve as a function of relative humidity showing deliquescent
growth of ammonium sulfate [(NH4)2 SO4] particles at the deliquescent point
(A, about 80% relative humidity [RH]), reversible hygroscopic growth of
ammonium sulfate solution droplets at RH greater than 80%, and hysteresis
(the droplet remains supersaturated as the RH decreases below 80%) until
the crystallization point (B, about 38% RH) is reached.
Source: Adapted from National Research Council (1993) and Tang (1980).
1 Water growth behavior for hygroscopic materials commonly found in atmospheric aerosol
2 in pure form or in some mixtures is generally well known as a result of laboratory measurements
3 (Tang and Munkelwitz, 1994; Tang, 1997). Models that calculate water growth of mixtures from
April 2002
4-150
DRAFT-DO NOT QUOTE OR CITE
-------
1 known solubility properties of many common water-soluble chemicals have long been available
2 (Zdanovskii, 1948) and have been successfully applied to determine growth for particles with
3 known composition (Saxena and Peterson, 1981; Pilinis et al., 1995; Saxena et al., 1993).
4 The water growth of individual ambient particles can be directly measured using a
5 humidity-controlled tandem differential mobility analyzer or TDMA (McMurry and Stolzenburg,
6 1989; Zhang et al., 1993). Inferences can be made about the mixtures of soluble and insoluble
7 particle components by comparing TDMA measured growth and size-resolved aerosol
8 composition data with water growth model predictions (Pitchford and McMurry, 1994; Zhang
9 et al., 1993; Saxena et al., 1995). A practical limitation of TDMA measurements in investigating
10 aerosol optical properties is that particles with diameter greater than 0.5 //m are not well
11 measured by this approach.
12 Accounting for water growth of atmospheric aerosols is important for visibility because
13 particles containing hygroscopic or deliquescent materials change size and index of refraction,
14 and hence scattering efficiency, with changing relative humidity. The nonlinear nature of particle
15 growth curves for hygroscopic aerosols means that substantial light scattering changes result
16 from modest relative humidity changes under humid conditions (relative humidity > 90%). The
17 magnitude of the water growth effect on light scattering for ambient aerosols can be directly
18 measured with humidity-controlled nephelometer measurements (Day et al., 1997).
19 Measurements of water growth effects on scattering are compared to results of water growth and
20 Mie scattering models applied to size-resolved composition data using various mixture
21 assumptions to infer average mixture and other aerosol characteristics (Malm et al., 2000).
22 While the importance of inorganic hygroscopic particles is well understood, the role of
23 organic compounds in particle water growth has been the subject of recent investigations.
24 In their interpretation of TDMA and particle composition data from two locations, Saxena et al.
25 (1995) made the case that organic components of the aerosol enhanced water absorption by
26 particles at a remote desert location and retarded water absorption at an urban location. They
27 speculated that the latter might be due to hydrophobic organic material coatings on inorganic
28 hygroscopic particles.
29 While some of the thousands of organic compounds that are in atmospheric aerosols are
30 known to be hygroscopic and while a significant fraction of the organic aerosol material is
31 known to be water soluble, there is a general lack of water absorption data for most organic
April 2002 4-151 DRAFT-DO NOT QUOTE OR CITE
-------
1 compounds. The incomplete water solubility data, combined with incomplete data on the
2 abundance of the numerous organic compounds in ambient aerosols, means that organic water
3 growth model calculations are not a reasonable approach to assessing the importance of water
4 growth by organic aerosol components in the atmosphere. To overcome this constraint, Saxena
5 et al. (1995) compared organic concentration to the difference between total aerosol water
6 measured by TDMA and model-estimated water for the inorganic hygroscopic aerosol
7 components. One the other hand, Pitchford and McMurry (1994) using the same remote location
8 data set showed that on six of the eight sampling days water uptake by the sulfates and nitrates
9 could account for all of the measured water absorption.
10 Swietlicki et al. (1999) made TDMA measurements in northern England and found that
11 growth takes place in two modes, one mode being less hygroscopic that the other. They
12 concluded that growth could be attributed to the inorganic content of the aerosol. Cocker et al.
13 (2001) measured hygroscopic properties of Pasadena, California aerosol and concluded that
14 growth factors increased when forest fires were present. McDow et al. (1995) measured water
15 uptake by diesel soot, automobile exhaust, and wood smoke particles. They found all three
16 emission types absorbed water, with the wood smoke sample weight increasing by about 10% as
17 sample relative humidity was increased; whereas diesel soot sample weight increased by only 2%
18 to 3%. Chughtai et al. (1999) examined hydration characteristics of a number of anthropogenic
19 and natural organic materials. They found surface water adsorption to increase with age and
20 surface oxidation. Analysis of humidity controlled and size-resolved chemistry data from Great
21 Smoky Mountains and Grand Canyon National Parks (Malm et al., 1997; Malm and
22 Kreidenweis, 1996; Malm et al., 2000) show that, to within the measurement uncertainty and
23 modeling assumptions, ambient organic aerosol are at most weakly hygroscopic.
24 A more detailed discussion of the effects of relative humidity on the size distribution of
25 ambient particles appears in Chapter 2 of this document.
26
27 4.3.3 Relationships Between Particles and Visibility
28 Visibility, referring to the appearance of scenic elements in an observer's line of sight,
29 depends on more than the optical characteristics of the atmosphere. Numerous scene and lighting
30 characteristics are important to this broad definition of visibility. However, under a variety of
April 2002 4-152 DRAFT-DO NOT QUOTE OR CITE
-------
1 viewing conditions, visibility reduction or haziness is directly related to the extinction
2 coefficient, which, as shown above, is related to the concentrations of ambient particles.
3 Often visibility conditions are communicated in terms of the visual range, which is
4 commonly taken to be the greatest distance that a large dark object (e.g., a mountain in shadow)
5 can be seen against the background sky (Middleton, 1952). Visual range was developed and
6 continues to function well as an aid in military operations and transportation safety. An inverse
7 relationship between visual range and the light extinction coefficient, know as the Koschmeider
8 constant, was developed using a number of restrictive assumptions about lighting, scene, and
9 atmospheric conditions.
10
11 Visual Range = 3.912/bext (4-7)
12
13 where visual range is in kilometers, &ext is in km"1, and a threshold contrast of 2% is assumed.
14 If bext is in Mm"1, the Koschmeider constant becomes 3,912.
15 A new index of haziness, expressed in deciview (dv) units, is also very simply related to the
16 light extinction coefficient (Pitchford and Malm, 1994).
17
18 Haziness (dv) = 10 l^bJlOMm1) (4-8)
19
20 An important characteristic of this visibility index is that it is more nearly linearly related to
21 perceived changes in haze level than either visual range or light extinction. A change of 1 or
22 2 dv in uniform haze under many viewing conditions will be seen as a small but noticeable
23 change in the appearance of a scene regardless of the initial haze condition.
24 Figure 4-27 illustrates the relationship of light extinction in Mm"1, deciview index, and
25 visual range in kilometers. Although the deciview is related to extinction, it is scaled in such a
26 way that is perceptually correct (Fox et al., 1999).
27 Comparisons of paired light extinction coefficient (or scattering coefficient) and particle
28 mass concentration data reveal a definite but noisy linear relationship. In general such a
29 relationship can be improved by either restricting the data to periods of low relative humidity or
30 by empirically adjusting for the nonlinear effects of water growth using relative humidity data
31 (White and Roberts, 1977; Malm, 1989). Where particle speciation data for the major aerosol
April 2002 4-153 DRAFT-DO NOT QUOTE OR CITE
-------
Extinction (Mm"1)
Deci views (dv)
10 20 30 40 50 70 100 200 300 400 500 700 1000
I I I
I
0
I
I
7
I
I
11
I
I
14
I
I
16
I
I Mil
19 23
i mi
i
30
I
I
34
I
I
37
I
I
39
I
I I III
42 46
I I III
Visual Range (km) 40o 2oo 130 100 so eo 40 20 13 10 s 64
Figure 4-27. Comparison of extinction (Mm"1) and visual range (km).
Source: Fox et al. (1999).
1 components are available, the relationship between particles and light extinction can be further
2 improved by treating the individual major components separately.
3 Most routine aerosol monitoring programs and many special study visibility
4 characterization programs were designed to measure the major aerosol components (Malm et al.,
5 1994; Tombach and Thurston, 1994; Watson et al., 1990); they were not designed to determine
6 the microphysical and chemical characteristics of these species. However, the inherent
7 limitations of estimating aerosol optical properties from bulk aerosol measurements have been
8 addressed, at least in part, by a number of authors. For instance, Ouimette and Flagan (1982)
9 have shown from basic theoretical considerations that if an aerosol is mixed externally (i.e.,
10 separate particles contain the major aerosol components), or if in an internally mixed aerosol the
11 index of refraction is not a function of composition or size and the aerosol density is independent
12 of volume, then
13
14 bsp=Zafni (4-9)
15
16 where a; is the specific mass scattering efficiency and m; is the mass of the individual aerosol
17 species.
18 Sloane (1983, 1984, 1986), Sloane and Wolff (1985), and more recently Lowenthal et al.
19 (1995) and Malm and Kreidenweiss (1997) have shown that differences in estimated specific
20 scattering between external and internal model assumptions are usually less than about 10%.
21 In the absence of detailed microphysical and chemical information of ambient particles, the
April 2002 4-154 DRAFT-DO NOT QUOTE OR CITE
-------
1 above studies demonstrate that a reasonable estimate of aerosol extinction can be achieved by
2 assuming each species is externally mixed.
3 The latest IMPROVE Program report (Malm, 2000) includes calculated aerosol light
4 extinction for each of the five major fine fraction particle (PM2 5) components plus coarse
5 fraction mass (PM10_25) and sums them for an estimate of total light extinction in Mm"1 using the
6 following algorithm:
7
b«t = (l)f(RH) [SULFATE] +
(3)f(RH) [NITRATE]
+(4) [ORGANIC CARBON]
+( 1 0) [LIGHT ABSORBING CARBON] (4- 1 0)
+(1) [SOIL]
+(0.6) [COARSE PMJ
+10 (for Rayleigh scattering by gases)
1 where each PM term is the product of a constant dry extinction efficiency for that species, the
2 mass concentration of the species, and, for sulfate and nitrate, an adjustment factor that is a
3 function of relative humidity to account for their hygroscopic behavior. The relative humidity
4 adjustment term for sulfate and nitrate, shown in Figure 4-28, is based upon the ammonium
5 sulfate growth curve, shown in Figure 4-26, smoothed between the upper and lower curves of the
6 hysteresis loop for the relative humidity range of 30-80%.
7 The extinction efficiencies for soil and coarse mass used in this algorithm are taken from a
8 literature review by Trijonis et al. (1987). The extinction efficiency for light absorbing
9 (elemental) carbon of 10 m2/g is consistent with the value reported by Moosmiiller et al. (1998)
10 corresponding to A = 0.53 in the middle of the visible light spectrum. The dry extinction
1 1 efficiencies of 3 m2/g for sulfate and nitrate species and 4 m2/g for organic species are based on
12 literature reviews by Trijonis et al. (1991) and by White (1991). Trijonis' best estimate for
1 3 sulfates is 2. 5 m2/g with an uncertainty of a factor of 2, while White' s average low and high
14 estimates for the rural West are 3.0 and 3.7 m2/g, respectively. For organics Trijonis estimates a
1 5 dry extinction efficiency of 3 .75 m2/g with an uncertainty of a factor of 2, and White' s range for
April 2002 4-155 DRAFT-DO NOT QUOTE OR CITE
-------
I
a:
o
"o
ro
u_
O)
!E
D)
30 40 50 60 70 80 90 100
Relative Humidity (percent)
Figure 4-28. Relative humidity adjustment factor, f(RH), for ammonium sulfate as a
function of relative humidity.
Source: Malm et al. (2000).
1 the rural West is 1.8 to 4.1 m2/g. Malm et al. (1996) and Malm (2000) used this algorithm to
2 successfully reconstruct scattering at a total of eleven IMPROVE monitoring sites.
3 Malm (2000) used additional sophisticated aerosol size, composition, and microphysical
4 data from a special study at the Great Smoky Mountain National Park to compare the
5 performance of a number of models for calculating light extinction. He found that the simplist
6 approach adequately predicted for periods of low light scattering but under-predicted by about
7 30% during periods of high sulfate concentration. The greatest improvement over the simple
8 model was obtained by including the degree of sulfate ammoniation in the model, which
9 produced better estimates of extinction coefficient over the entire range.
10
11 4.3.4 Photographic Modeling of Visibility Impairment
12 None of the visibility indices communicate visibility associated with various aerosol
13 conditions as well as directly seeing their effects on a scene. However, photographic modeling
April 2002
4-156
DRAFT-DO NOT QUOTE OR CITE
-------
1 for the representation of haze can be useful in portraying changes in visibility specifically due to
2 changes in air pollutant concentrations. Photographic modeling holds constant the effects of sun
3 angle, cloud cover, and relative humidity and is a cost-effective method of evaluating various air
4 quality scenarios. This is difficult to do with actual photographs because of the range of possible
5 conditions in the same scene over multiple days; and, over time, photographs can be expensive to
6 produce. One of the limitations in using photographic models for representation of haze is that
7 haze is assumed to be uniformly distributed throughout the scene and selected conditions are
8 idealized, so the full range of conditions that occur in a scene are not represented.
9 Eldering et al. (1996) proposed the use of a model that uses simulated photographs from
10 satellite and topographic images to evaluate the effect of atmospheric aerosols and gases on
11 visibility. Use of this model requires ground-based photography and size distribution and
12 chemical composition of atmospheric aerosols, NO2 concentration, temperature, and relative
13 humidity for a clear day, for comparison purposes. Light extinction and sky color are then
14 calculated based on differences in aerosol size distribution, NO2 concentration, temperature, and
15 relative humidity. The images created represent natural landscape elements.
16 Molenar et al. (1994) provides a discussion of existing visual air quality simulation
17 methods based on techniques under development for the past 20 years. A photograph taken on a
18 very clean, cloud-free day serves as the base image. The photograph is taken during the season
19 and at the same time of day as the scene to be modeled. The light extinction represented by the
20 scene is derived from aerosol and optical data associated with the day the image was taken, or it
21 is estimated from contrast measurements of features in the image. The image is then digitized to
22 assign an optical density to each picture element (pixel) for the wavelength bands of interest.
23 A detailed topographic map and an interactive image-processing display system is used to
24 determine the specific distance, elevation angle, and azimuth angle for each element in the
25 picture with respect to the observer's position.
26 Various models are employed to allow the presentation of different air quality scenarios.
27 The output from atmospheric aerosol models (e.g., extinction, scattering coefficients, single
28 scattering albedo, and scattering phase matrix) is incorporated into radiative transfer models to
29 calculate the changes in radiant energy (path radiance, image radiance, sky radiance, terrain
30 radiance) caused by scattering and absorption by gases and particles as it passes through the
April 2002 4-157 DRAFT-DO NOT QUOTE OR CITE
-------
1 atmosphere. Atmospheric aerosol models are also use to model the effect of relative humidity on
2 the visual air quality (Molenar et al., 1994).
3 Molenar et al. (1994) has developed a system call WinHaze that permits the viewing of
4 computer-generated uniform hazes superimposed on digitized scenic photographs of both remote
5 and urban scenes. The program simulates changes in visual air quality imagery from user-
6 specified changes in optical parameters (e.g., aext, visual range, or deciview values) or aerosol
7 species concentrations. WinHaze includes imaging for various Class I national parks and
8 wilderness areas and Boston, MA; Dallas, TX; Denver, CO; Fort Collins, CO; Phoenix, AZ;
9 and Tucson, AZ. The computer software is available through the IMPROVE website
10 (http://vista.cira.colostate.edu/ improve/).
11
12 4.3.5 Visibility Monitoring Methods and Networks
13 Visibility monitoring studies measure the properties of the atmosphere either at the sampler
14 inlets (point measurements), as is the case with air quality measurements, or by determining the
15 optical properties of a sight path through the atmosphere (path measurements). Instrumental
16 methods for measuring visibility are generally of three types: (1) direct measurement of light
17 extinction of a sight path using a transmissometer, (2) measurement of light scattering at one
18 location using an integrating nephelometer, and (3) measurement of ambient aerosol mass
19 concentration and composition (Mathai, 1995).
20 The largest instrumental visibility monitoring network in the United States is the
21 Automated Surface Observing System (ASOS). The Automated Surface Observing System has
22 been commissioned by the National Weather Service, Federal Aviation Administration, and
23 Department of Defense at more than 900 airports. The system is designed to objectively measure
24 the clarity of the air versus the more subjective evaluations of human observations. The system
25 provides real-time data for airport visibility.
26 The visibility sensor, instead of measuring how far one can see, measures the clarity of the
27 air using a forward scatter visibility meter. The clarity is then converted to what would be
28 perceived by the human eye using a value called Sensor Equivalent Visibility (SEV). Values
29 derived from the sensor are not affected by terrain, location, buildings, trees, lights, or cloud
30 layers near the surface. The amount of moisture, dust, snow, rain, and particles in the light beam
31 will affect the amount of light scattered. The sensor transmits 1-min values based on rolling
April 2002 4-158 DRAFT-DO NOT QUOTE OR CITE
-------
1 10-min periods. The value provides a generally accurate and representative visibility
2 measurement within a 2 to 3 mile radius of the site. The forward scatter meter was found to
3 correlate fairly well with extinction coefficient measurements from the Optec Transmissometer
4 (National Weather Service, 1998).
5 Visibility data from the ASOS network is reported in terms of visual range in increments of
6 1/4 to 1 statute mile. Visual range conditions exceeding 10 miles are truncated to 10 miles for
7 real-time reporting purposes. Data is not extensively archived at ASOS locations. However,
8 researchers have been able to download the raw data directly from certain sites. In addition,
9 since 1998, the raw visibility data (including light extinction measurements corresponding to
10 visual ranges exceeding 10 miles) have been archived for a number of sites. These data are
11 available from the National Climatic Data Center.
12 The ASOS data may be useful for aiding in the characterization of visibility conditions in
13 urban and suburban areas across the country. It also may be useful in future analyses to better
14 understand the effects of fine PM on visibility in Class I areas. The Agency is in the process of
15 analyzing ASOS data for a limited number of sites to determine how well it correlates to
16 particulate matter monitoring results. In addition, the analysis will evaluate annual averages and
17 seasonal, monthly, and daily visibility conditions (U.S. Environmental Protection Agency, 2001).
18 The Agency expects that the results of these analyses will be available for inclusion in the final
19 PMAQCD.
20 The largest monitoring network that includes both visibility and aerosol conditions is the
21 Interagency Monitoring of Protected Visual Environments (IMPROVE) network. This network
22 was formed in 1987 as a collaborative effort between Federal, regional, and state organizations
23 responsible for protection of visibility in the 156 mandatory Class I Federal areas (national parks
24 and wilderness areas) and other areas of interest to land management agencies, states, tribes, and
25 other organizations (National Park Service, 1998; U.S. Environmental Protection Agency, 1996a;
26 U.S. Environmental Protection Agency, 1995b; Eldred et al., 1997; Perry et al., 1997; Sisler and
27 Malm, 2000; U.S. Environmental Protection Agency, 1999b). It is predominantly a rural-based
28 network, with more than 140 sites across the country. The primary monitoring objectives of the
29 IMPROVE program are to document current visibility conditions in the mandatory Class I areas,
30 identify anthropogenic chemical species and emission sources of visibility impairment through
31 the collection of speciated PM2 5 data, and document long-term trends for assessing progress
April 2002 4-159 DRAFT-DO NOT QUOTE OR CITE
-------
1 towards elimination of anthropogenic visibility impairment. The IMPROVE network has also
2 been involved in visibility related research, including the advancement of visibility monitoring
3 instrumentation and analysis techniques and visibility monitoring and source attribution field
4 studies (National Park Service, 1998; Evans and Pitchford, 1991).
5 Visibility monitoring under the IMPROVE network can be divided into three categories:
6 aerosol, optical, and scene. Twenty-four hour PM2 5 and PM10 aerosol samples are collected at
7 least every third day utilizing filter-based aerosol technology. The PM2 5 samples are analyzed to
8 determined the mass concentration of the major parti culate constituents (sulfates, nitrates,
9 organic carbon compounds, elemental carbon, chlorides, and crustal elements) and for elements
10 that indicate sources of visibility-impairing particles (trace elements and ions). Optical
11 monitoring provides a direct measurement of light scattering and absorption. Color photographic
12 imaging documents the appearance of the scene under a variety of air quality and illumination
13 conditions (U.S. Environmental Protection Agency, 1999b). It is anticipated that all data
14 generated by the IMPROVE network will be added to the AIRS database.
15 The U.S. Environmental Protection Agency has deployed a new national monitoring
16 network (Federal Reference Method Monitoring network) designed to assess PM2 5
17 concentrations and composition. As of early 2001, 1,108 monitoring sites were in operation
18 (including more than 250 urban sites) and 1,044 sites had reported data to the Aerometric
19 Informational Retrieval System (AIRS). Analyses of these data are expected to provide a more
20 complete understanding of visibility conditions, in particular urban visibility, across the country.
21 The PM25 monitoring effort has been coordinated with visibility monitoring efforts currently in
22 place to maximize the benefits of all of the monitoring programs (U.S. Environmental Protection
23 Agency, 1997b; U.S. Environmental Protection Agency, 2000b; U. S. Environmental Protection
24 Agency, 2001).
25 The Northeast States for Coordinated Air Use Management (NESCAUM) has established a
26 real-time visibility monitoring network (CAMNET) using digital photographic imaging.
27 Currently, there are digital photographic imaging for five urban locations (Boston, MA;
28 Burlington, VT; Hartford, CT; Newark, NJ; and New York City, NY), and two rural locations
29 (Acadia National Park, ME and Mt. Washington, NH). The visibility images are updated every
30 15 minutes. Near real-time air pollution and meteorological data are updated every hour.
31 Archived images will be available to understand the visual effects of parti culate matter air
April 2002 4-160 DRAFT-DO NOT QUOTE OR CITE
-------
1 pollution in the Northeast. CAMNET may be accessed at www.hazecam.net (Northeast States
2 for Coordinated Air Use Management, 2002; Leslie, 2001).
3
4 4.3.6 Visibility Impairment: Trends and Current Conditions
5 In the United States, visibility impairment is caused by particles primarily composed of
6 sulfates, nitrates, organic compounds, carbon soot, and crustal dust. Generally, sulfate is the
7 major component responsible for visibility impairment in the eastern United States. However,
8 nitrates, organic compounds, carbon soot, and crustal material are significant contributors to
9 visibility impairment in some locations (Sisler and Cahill, 1993).
10
11 Trends in Visibility Impairment
12 Trends in visibility impairment or haziness often are used as indicators of trends in fine
13 particles. Observations of visual range, obtained by the National Weather Service and available
14 through the National Climatic Data Center of the National Oceanic and Atmospheric
15 Administration, provide one of the few truly long-term, daily records of impairment related to air
16 pollution. After some manipulation, including correction for relative humidity effects, the visual
17 range data can be used as an indicator of fine mode particle pollution. The data reduction
18 process and analyses of resulting trends have been reported by Schichtel et al. (2001), Husar et al.
19 (1994), Husar and Wilson (1993), and Husar et al. (1981).
20 There are many statistical approaches to estimating trends. These approaches include
21 simple correlation and regression analyses, time series analyses, and methods based on
22 non-parametric statistics. A discussion and comparison of the methods for the detection of linear
23 trends is provided in Hess et al. (2001). Schimek (1981) provides a discussion of nonlinear
24 trends. In its annual air quality trends report, the U. S. Environmental Protection Agency
25 characterized trends using a non-parametric regression analysis approach commonly referred to
26 as the Theil test (U.S. Environmental Protection Agency, 1998; Hollander and Wolfe, 1973).
27 Generally, visibility impairment is greatest in the eastern United States and southern
28 California. Visibility impairment or haziness in the southeastern United States, caused largely by
29 sulfate formed from SO2, is greatest in the humid summer months because of the ability of
30 sulfate to absorb atmospheric water vapor, followed by the spring and fall, and winter. Summer
31 haziness increased in the southeastern United States from the 1950s to 1980 along with
April 2002 4-161 DRAFT-DO NOT QUOTE OR CITE
-------
1 increasing SO2 emissions. A decrease in haziness between 1980 and 1995 corresponded with a
2 similar decrease in sulfur emissions (Schichtel et al., 2001). A statistically significant increase in
3 summer sulfate concentrations was noted in two Class I areas in the eastern United States
4 (Shenandoah and the Great Smoky Mountains) from 1982 to 1992 (Eldred et al., 1993; Cahill
5 et al., 1996). During that time period, the majority of the Southwest showed decreasing sulfur
6 (Eldred et al., 1993; Eldred and Cahill, 1994). The increasing sulfate concentrations were later
7 shown to correlate with an increased trend in hazy days at those two locations (Iyer et al., 2000).
8 The U.S. Environmental Protection Agency's National Air Quality and Emission Trends
9 Report summarized the regional trends and current conditions in 35 Class I areas and one urban
10 area (Washington, DC) using data from the IMPROVE network (U.S. Environmental Protection
11 Agency, 2001). The visibility trends analysis is an aggregate of 10 eastern Class I areas and
12 26 western Class I areas. Trends were presented for annual average values for the clearest
13 ("best") 20% , middle ("typical") 20%, and haziest ("worst") 20% of the days monitored each
14 year. The visibility trends, given in changes in deciview values, for the eastern and western sites
15 are illustrated in Figures 4-29a and 4-29b. From the figures it can be seen that the haziest days in
16 the West are equivalent to the best days in the East. In the East there was a 16% (1.5 deciview)
17 improvement in haziness on the clearest days since 1992. Improvements in visibility were noted
18 in the East for the haziest days; however, based on monitoring data for 1999, visibility remains
19 significantly impaired with a visual range of 23 km for the haziest days compared to a mean
20 visual range of 84 km for the clearest days. A 25 and!4% improvement in visibility impairment
21 was seen for the clearest and middle days in the West, respectively. Conditions for the haziest
22 days degraded by 18.5% (1.7 deciviews) between 1997 and 1999, but were relatively unchanged
23 compared to 1990 conditions (U.S. Environmental Protection Agency, 2001).
24 Figures 4-3Oa and 4-3Ob illustrate aggregate trends in aerosol light extinction, including
25 trends by major aerosol component for the haziest 20% of days monitored for the 10 eastern
26 Class I areas from 1992 to 1999 and the haziest 20% of days monitored for the 26 western Class I
27 areas from 1990 to 1999. The report also includes a number of maps characterizing aerosol light
28 extinction and key components at 36 IMPROVE sites (all rural except Washington, DC) for 1997
29 through 1999 (U.S. Environmental Protection Agency, 2001).
April 2002 4-162 DRAFT-DO NOT QUOTE OR CITE
-------
Visibility Impairment (deciviews)
35
30
25
20
15
10
5
n
Haziest 20-percent
' -
Typical 20-percent
Clearest 20-percent
92 93 94 95 96 97 98 99
Year Measured
Figure 4-29a. Aggregate visibility trends (in deciviews) for 10 eastern Class 1 areas.
ility Impairment (deciviews)
_Q
(/>
>
GO
30
25
20
15
10
5
n
Haziest 20-percent
Typical 20-percent
Clearest 20-percent
90 91 92 93 94 95 96 97 98 99
Year Measured
Figure 4-29b. Aggregate visibility trends (in deciviews) for 26 western Class 1 areas.
Source: U.S. Environmental Protection Agency (2001).
April 2002
4-163
DRAFT-DO NOT QUOTE OR CITE
-------
tz
o
200
150
uj 100
xi
_ 50
o
co
O
CD
0
92 93 94 95 96 97
Year Measured
98 99
H Organic Carbon
• Nitrate
Elemental Carbon
U Crystal Material
Figure 4-30a. Eastern class I area aggregate trends in aerosol light extinction on the 20%
haziest days, including trends by major aerosol component.
E
£=
_o
"»8™rf
O
X
UJ
^™rf
.c
O)
o
CO
o
100
75
50
25
0
90 91 92 93 94 95 96 97 98 99
Year Measured
^Organic Carbon
• Nitrate
ySuIfate
Elemental Carbon
Crustal Material
Figure 4-30b. Western class I area aggregate trends in aerosol light extinction on the 20%
haziest days, including trends by major aerosol component.
Source: U.S. Environmental Protection Agency (2001).
April 2002
4-164
DRAFT-DO NOT QUOTE OR CITE
-------
1 Current Conditions
2 Current visibility conditions have been well-characterized for Class I areas using updated
3 data from the IMPROVE network (U.S. Environmental Protection Agency, 2001; Malm et al.,
4 2000; IMPROVE, 1998). During recent decades, daytime visibility conditions at all major
5 airports throughout the United States were recorded hourly by human observation. These data
6 were used to determine current visibility conditions and visibility trends in the United States, as
7 well as the spatial distribution of visibility conditions (Trijonis et al., 1991). The use of airport
8 human observation is being replaced by an automated observing system, Automated Surface
9 Observing System (ASOS). More than 900 airports are currently commissioned. Additionally,
10 the U.S. Environmental Protection Agency has deployed a new national monitoring network to
11 assess PM2 5 concentrations and composition. Visibility conditions for urban and suburban areas
12 will become more widely available as data from the national PM2 5 speciation monitoring
13 network and the ASOS airport visibility network are further analyzed.
14
15
16 4.4 EFFECTS ON MATERIALS
17 Effects of air pollution on materials are related to both aesthetic appeal and physical
18 damage. Studies have demonstrated that particles, primarily consisting of carbonaceous
19 compounds, cause soiling of commonly used building materials and culturally important items,
20 such as statutes and works of art. Physical damage from the dry deposition of air pollutants, such
21 as PM (especially sulfates and nitrates) and SO2, and the absorption or adsorption of corrosive
22 agents on deposited particles also can result in the acceleration of naturally occurring weathering
23 processes of man-made building and cultural materials.
24 In the atmosphere, PM may be "primary," existing in the same form in which it was
25 emitted, or "secondary," formed by the chemical reactions of free, absorbed, or dissolved gases.
26 The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen ions;
27 particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
28 material. A substantial fraction of the fine particle mass, particularly during the warmer months,
29 is secondary sulfate and nitrate. Sulfates may be formed by the gas-phase conversion of SO2 to
30 H2SO4 by OH radicals and aqueous-phase reactions of SO2 with H2O2, O3, or O2. During the day,
31 NO2 may be converted to nitric acid (HNO3) by reacting with OH radicals. Nitrogen dioxide also
April 2002 4-165 DRAFT-DO NOT QUOTE OR CITE
-------
1 can be oxidized to HNO3 by a sequence of reactions initiated by O3. A more detailed discussion
2 of the atmospheric chemistry of PM appears in Chapter 2 of this document.
3
4 4.4.1 Corrosive Effects of Particles and Sulfur Dioxide on
5 Man-Made Surfaces
6 Limited new studies have been published that better define the role of air pollution in
7 materials damage. This section briefly summarizes information on exposure particle related
8 effects on materials and sulfur-containing pollutants (formed by the chemical reactions of SO2
9 with other atmospheric pollutants) addressed in the 1996 PM AQCD (U.S. Environmental
10 Protection Agency, 1996a) and presents relevant information published since completion of that
11 document. The effects of nitrates on man-made building materials and naturally occurring
12 cultural materials was discussed in the earlier EPA Nitrogen Oxides Criteria Document (U.S.
13 Environmental Protection Agency, 1993).
14
15 4.4.1.1 Metals
16 Metals undergo natural weathering processes in the absence of environmental pollutants.
17 The additive effect of pollutants on the natural weathering processes depend on the nature of the
18 pollutant, and the deposition rate (the uptake of a pollutant by the material's surface), and the
19 presence of moisture. The influence of the metal-protective corrosion film, the presence of other
20 surface electrolytes, the orientation of the metal surface, the presence of surface moisture, and the
21 variability in the electrochemical reactions will also contribute to the effect of pollutant exposure
22 on metal surfaces.
23 Several studies demonstrate the importance of the duration of surface wetness (caused by
24 dew and fog condensation and rain) on metals. Surface moisture facilitates the deposition of
25 pollutants, especially SO2, and promotes corrosive electrochemical reactions on metals (Haynie
26 and Upham, 1974; Sydberger and Ericsson, 1977). Of critical importance is the formation of
27 hygroscopic salts on the metal that increases the duration of surface wetness and, thereby,
28 enhances the corrosion process.
29 Pitchford and McMurry (1994) and Zhang et al. (1993) demonstrated particle size-related
30 effects of relative humidity. The effect of temperature on the rate of corrosion is complex.
31 Under normal temperature conditions, temperature would not have an affect on the rate of
April 2002 4-166 DRAFT-DO NOT QUOTE OR CITE
-------
1 corrosion; but when the temperature decreases, the relative humidity increases and the diffusivity
2 decreases. The corrosion rate decreases as the temperature approaches freezing because ice
3 prohibits the diffusion of SO2 to the metal surface and minimizes electrochemical processes
4 (Haynie, 1980; Biefer, 1981; Sereda, 1974).
5 The metal protective corrosion film (i.e., the rust layer on metal surfaces) provides some
6 protection against further corrosion. The effectiveness of the corrosion film in slowing down the
7 corrosion process is affected by the solubility of the corrosion layer and the concentration and
8 deposition rate of pollutants. If the metal-protective corrosion film is insoluble, it may add some
9 protection against acidic pollutants. An atmospheric corrosion model that considers the
10 formation and dissolution of the corrosion film on galvanized steel was proposed by Spence et al.
11 (1992). The model considers the effects of SO2, rain acidity, and the time of wetness on the rate
12 of corrosion. Although the model does not specifically characterize particle effects, the
13 contribution of particulate sulfate was considered in model development.
14 Whether suspended particles actually impact on the corrosion of metals is not clear.
15 Several studies suggest that suspended particles will promote the corrosion of metals (Goodwin
16 et al., 1969; Barton, 1958; Sanyal and Singhania, 1956; Baedecker et al., 1991); however, other
17 studies have not demonstrated a correlation between particle exposure and metal corrosion
18 (Mansfeld, 1980; Edney et al., 1989). Walton et al. (1982) suggested that catalytic species within
19 several species in fly ash promote the oxidation of SOX to a corrosive state. Still other
20 researchers indicate that the catalytic effect of particles is not significant and that the corrosion
21 rate is dependent on the conductance of the thin-film surface electrolytes during periods of
22 wetness. Soluble particles likely increase the solution conductance (Skerry et al., 1988; Askey
23 etal., 1993).
24 The corrosion of most ferrous metals (iron, steel, and steel alloys) is increased by
25 increasing SO2 exposure. Steels are susceptible to corrosion when exposed to SO2 in the absence
26 of protective organic or metallic coatings. Studies on the corrosive effects of SO2 on steel
27 indicate that the rate of corrosion increases with increasing SO2 and is dependent on the
28 deposition rate of the SO2 (Baedecker et al., 1991; Butlin et al., 1992a). The corrosive effects of
29 SO2 on aluminum is exposure-dependent, but appears to be insignificant (Haynie, 1976; Fink
30 et al., 1971; Butlin et al., 1992a). The rate of formation of the patina (protective covering) on
31 copper can take as long as 5 years and is dependent on the SO2 concentration, deposition rate,
April 2002 4-167 DRAFT-DO NOT QUOTE OR CITE
-------
1 temperature, and relative humidity (Simpson and Horrobin, 1970). Further corrosion is
2 controlled by the availability of copper to react with deposited pollutants (Graedel et al., 1987).
3 Butlin et al. (1992a), Baedecker et al. (1991), and Cramer et al. (1989) reported an average
4 corrosion rate of 1 //m/year for copper; however, less than a third of the corrosion was attributed
5 to SO2 exposure, suggesting that the rate of patina formation was more dependent on factors
6 other than SO2. A recent report by Strandberg and Johansson (1997) showed relative humidity to
7 be the primary factor in copper corrosion and patina formation. The results of the studies on
8 particles and SO2 corrosion of metals are summarized in Table 4-18.
9
10 4.4.1.2 Painted Finishes
11 Exposure to air pollutants affect the durability of paint finishes by promoting discoloration,
12 chalking, loss of gloss, erosion, blistering, and peeling. Evidence exists that indicates particles
13 can damage painted finishes by serving as carriers for corrosive pollutants (Cowling and Roberts,
14 1954) or by staining and pitting of the painted surfaces (Fochtman and Langer, 1957; Wolff et al.,
15 1990).
16 The erosion rate of oil-based house paint has been reported to be enhanced by exposure to
17 SO2 and high humidity. In a study by Spence et al. (1975), an erosion rate of 36.71 ±
18 8.03 //m/year was noted for oil-based house paint samples exposed to SO2 (78.6 //g/m3), O3
19 (156.8 //g/m3), and NO2 (94 //g/m3) and low humidity (50%). The erosion rate increased with
20 increased SO2 and humidity. The authors concluded that SO2 and humidity accounted for 61% of
21 the erosion. Acrylic coil coating and vinyl coil coating shows less pollutant-related erosion.
22 Erosion rates range from 0.7 to 1.3 //m/year and 1.4 to 5.3 //m/year, respectively. Similar
23 findings on SO2-related erosion of oil-based house paints and coil coatings have been reported by
24 other researchers (Davis et al., 1990; Yocom and Grappone, 1976; Yocom and Upham, 1977;
25 Campbell et al., 1974). Several studies suggest that the effect of SO2 is caused by its reaction
26 with extender pigments such as calcium carbonate and zinc oxide (Campbell et al., 1974; Xu and
27 Balik, 1989; Edney, 1989; Edney et al., 1988, 1989). However, Miller et al. (1992) suggested
28 that calcium carbonate acts to protect paint substrates. Another study indicated that exposure to
29 SO2 can increase the drying time of some paints by reacting with certain drying oils and will
30 compete with the auto-oxidative curing mechanism responsible for crosslinking the binder
31 (Holbrow, 1962).
April 2002 4-168 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-18. CORROSIVE EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON METALS
to
o
o
to
Metal
Exposure Conditions
Comments
Source
H
6
o
o
H
O
o
HH
H
W
Mild Steel
Galvanized Steel
Zinc
Zinc
Carbon Steel
Weathering Steel
Aluminum
Aluminum
Specimens exposed to SO2 and O3 under natural
and artificial conditions, and to NO2 under natural
conditions. SO2 concentrations ranged from 2.1 to
60 /j,g/m3. Annual average concentrations were
about 20 ^g/m3. Meteorological conditions were
unaltered. Specimens exposed at 29 sites for
2 years for mild steel and 1 y for galvanized steel.
Rolled zinc specimens exposed at various sites
around the country (rural, industrialized, marine)
for up to 20 years. Actual pollutant exposures not
reported.
Specimens exposed at 5 sites for 1 to 5 years.
Average SO2 concentrations ranged from 2 ± 4 to
15 ± 17 ppb (5.2 ± 10.4 to 39.3 ± 44.5
PM concentrations ranged from 14 to 60
Highest pollutant concentrations recorded at
1 year exposure site.
See Baedecker et al. (1991) above for exposure
conditions.
See Baedecker et al. (1991) above for exposure
conditions.
See Butlin et al. (1992a) above for exposure
conditions.
Steel corrosion was dependent on long-term SO2
exposure. The corrosion rate was about 50 ^m/year
for mild steel specimens for most industrial sites, but
ranged from 21 to 71 ^m/year. The corrosion rate
ranged from 1.45 to 4.25 ^m/year for galvanized
steel. The authors concluded that rainfall also may
have a significant effect on galvanized steel based on
a corrosion rate of 3.4 ^m/year seen at a very wet
site.
The highest corrosion rates were associated with
industrialized environments and marine environments
in direct contact with salt spray.
Average corrosion rate ranged from 0.63 to
1.33 jum/y. The highest corrosion was noted in the
most industrialized area. However, the corrosion
rates did not differ significant regardless of the SO2
concentration, suggesting that SO2 exposure may not
be the dominant factor in zinc corrosion.
Average corrosion rate for samples exposed for
5 years ranged from 6.6 to 12.8 ^m/year for carbon
steel and 3.7 to 5.0 ^m/year for weathering steel.
Highest corrosion rate noted for samples exposed for
1 year.
Corrosion rate was very low at all sites and ranged
from 0.036 to 0. 106 ^on/year.
Corrosion greater on the under side of specimens,
possibly because of lack of washoff and increased
PM in area. Maximum corrosion rate was
0.85 ^m/year. Pit depths of up to 72 /j.m were noted
after 2 years of exposure.
Butlin etal. (1992a)
Showak and Dunbar
(1982)
Baedecker etal. (1991)
Cramer etal. (1989)
Baedecker etal. (1991)
Cramer etal. (1989)
Baedecker et al. (1991)
Butlin etal. (1992a)
-------
TABLE 4-18 (cont'd). CORROSIVE EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON METALS
to
o
o
to
Metal
Exposure Conditions
Comments
Source
J^.
o
Copper
Copper
Copper
Copper
See Baedecker et al. (1991) above for exposure
conditions.
See Butlin et al. (1992a) above for exposure
conditions.
Specimens exposed to 4 to 69 ppb (10.4 to
180.7 ,wg/m3) and 1.0 ppm (2,618.7 ,wg/m3)
SO, for 20h at various relative humidities.
Specimens exposed artificially to 0.49 ± 0.01 ppm
(187 ± 3.8 Mg/m3) SO2 for 4 weeks at 70 and 90%
relative humidity.
Average corrosion rate for 3- and 5-year exposures
was about 1 Aim/year but the soluble portion was less
than a third of that which could be contributed to SO2
exposure. Dry deposition of SO2 was not as
important in patina formation as wet deposition of H+.
Majority of test sites showed a corrosion rate of
1 ± 0.2 ^m/year. The corrosion rate was
1.48 ^m/year at the site receiving the most rainfall.
The lowest corrosion rate, 0.66 ^m/year, was
associated with low rainfall, low SO2.
SO2 had no effect on copper when relative humidity
was <15%. Increasing relative humidity increases
patina formation in presence of trace SO2.
No SO2-related effects were noted on copper
specimens exposed to high SO2 regardless of the
percent relative humidity.
Corrosive effect of SO2 on copper increased with
increasing relative humidity.
Baedecker et al. (1991)
Butlin etal. (1992a)
Strandberg and
Johansson (1997)
Eriksson et al. (1993)
H
6
o
o
H
O
o
HH
H
W
-------
1 4.4.1.3 Stone and Concrete
2 Numerous studies suggest that air pollutants can enhance the natural weathering processes
3 on building stone. The development of crusts on stone monuments have been attributed to the
4 interaction of the stone's surface with sulfur-containing pollutants, wet or dry deposition of
5 atmospheric particles, and dry deposition of gypsum particles from the atmosphere. Because of a
6 greater porosity and specific surface, mortars have a greater potential for reacting with
7 environmental pollutants (Zappia et al., 1998). Details on these studies are discussed in
8 Table 4-19. The stones most susceptible to the deteriorating effects of sulfur-containing
9 pollutants are the calcareous stones (limestone, marble, and carbonated cement). Exposure-
10 related damage to building stones result from the formation of salts in the stone that are
11 subsequently washed away during rain events leaving the stone surface more susceptible to the
12 effects of pollutants. Dry deposition of sulfur-containing pollutants promotes the formation of
13 gypsum on the stone's surface. Gypsum is a gray to black crusty material comprised mainly of
14 calcium sulfate dihydrate from the reaction of calcium carbonate (calcite) in the stone with
15 atmospheric SO2 and moisture (relative humidities exceeding 65%). Approximately 99% of the
16 sulfur in gypsum is sulfate because of the sulfation process caused by the deposition of SO2
17 aerosol. Sulfites also are present in the gypsum layer as an intermediate product (Sabbioni et al.,
18 1996; Ghedini et al., 2000; Gobbi et al., 1998; Zappia et al., 1998). Gypsum is more soluble than
19 calcite and is known to form on limestone, sandstones, and marble when exposed to SO2.
20 Gypsum also has been reported to form on granite stone by replacing silicate minerals with
21 calcite (Schiavon et al., 1995). Gypsum occupies a larger volume than the original stone, causing
22 the stone's surface to become cracked and pitted. The rough surface serves as a site for
23 deposition of airborne particles.
24 The dark colored gypsum is caused by surface deposition of carbonaceous particles
25 (noncarbonate carbon) from combustion processes occurring in the area (Sabbioni, 1995;
26 Saiz-Jimenez, 1993; Ausset et al., 1998), trace metals contained in the stone, dust, and numerous
27 other anthropogenic pollutants. After analyzing damaged layers of several stone monuments,
28 Zappia et al. (1993) found that the dark-colored damaged surfaces contained 70% gypsum and
29 20% noncarbonate carbon. The lighter colored damaged layers were exposed to rain and
30 contained 1% gypsum and 4% noncarbonate carbon. It is assumed that rain removes reaction
31 products, permitting further pollutant attack of the stone monument, and likely redeposits some
April 2002 4-171 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-19. CORROSIVE EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON STONE
to
o
o
to
Stone
Exposure Conditions
Comments
Source
to
H
6
o
o
H
O
o
HH
H
W
Vermont marble
Marble sandstone
Limestone
Portland limestone
White Mansfield
dolomitic sandstone
Monk's Park
limestone
Sandstones (calcite
and noncalcite
stones)
Limestones
Sandstones
Marble
Granite
Basalt
Portland limestone
Massangis Jaune
Roche limestone
White Mansfield
dolomitic
Runoff water was analyzed from seven summer
storms. SO2 concentration stated to be low.
Analysis of runoff water for five slabs test
exposed to ambient conditions at a angle of
30° to horizontal.
Ambient air conditions. Exposure ranged from
70 to 1065 days. Averaged pollutant exposure
ranged from 1.4 to 20.4 ppb (3.7 to 53.4 Mg/m3)
SO2; 4.1 to 41.1 ppbNOx; 2.4 to 17.4 ppb (4.5 to
32.7 Mg/m3) NO2; 10.1 to 25.6 ppb (19.8 to
50.2,ug/m3)O3.
Experimental tablets exposed under sheltered and
unsheltered ambient air conditions. Exposure for
1 and 2 years.
Ambient air; low concentrations of sulfates, SO2.
and nitrates; RH sufficient to produce
condensation on stones rarely occurred.
Ambient air; urban and rural locations in
Mediterranean.
Samples exposed to SO2, NO2, and NO at 10
ppmv, both with and without O3 and under dry
(coming to equilibrium with the 84% RH) or
wetted with CO2-equilibrated deionized water
conditions. Exposure was for 30 days.
Between 10 to 50% of calcium in runoff water estimated from Schuster et al.
gypsum formation from dry deposition of SO2. (1994)
Pollutant exposure related erosion was primarily caused by dry Baedecker
deposition of SO2 and nitric acid between rain events and wet et al. (1992)
deposition of hydrogen ion. Recession estimates ranged from 15 to
30 ,um/year for marble and 25 to 45 /^m/year for limestone. A large
portion of the erosion results from the reaction of CO2 with the
calcium in the stone.
Increased stone weight loss with increased SO2. Rainfall did not Webb et al.
significantly affect stone degradation. Stone loss associated with SO2 (1992)
exposure estimated to be 24 ^m/year. Slight trend in decreasing
stone loss with increasing length of exposure.
Significant correlations existed between the mean annual SO2 Butlin et al.
concentration, rainfall volume, and hydrogen ion loading and the (1992b)
weight changes.
Insignificant differences in erosion rate found between calcite and Petuskey et al.
noncalcite sandstone. Moisture affected the rate of pollutant (1995)
deposition and enhanced susceptibility to pollutant related erosion.
Rain events given as primary factor affecting stone erosion. Pollutant
related erosion judged to be insignificant.
Crusts on stones were found to contain two layers; top layer, usually Garcia-Valles
black in color, composed of gypsum between 40 and 400 /jm thick. et al. (1998)
Innermost layer, ranging from brown to orange in color, primarily
consisted of calcite, between 10 and 600 /jm thick. Gypsum-rich
layer thought to be the result of sulphation of the calcitic layer by
atmospheric pollutants or dry or wet deposition of atmospheric dust.
In the absence of moisture, little reaction is seen. SO2 is oxidized to Haneef et al.
sulfates in the presence of moisture. The effect is enhanced in the (1993)
presence of O3. Massangis Jaune Roche limestone was the least
affected by the pollutant exposure. Crust lined pores of specimens
exposed to SO2.
-------
TABLE 4-19 (cont'd). CORROSIVE EFFECTS OF PARTICIPATE MATTER AND SULFUR DIOXIDE ON STONE
Stone
Exposure Conditions
Comments
Source
Monk's Park
Portland limestone
Carrara marble
Travertine
Tranistone
Carrara marble
Georgia marble
Carrara marble
Monk's Park
limestone
Portland limestone
Lime mortar
Pozzolan mortar
Cement mortar
Limestone
Travertine marble
Samples exposed for 2 mo under both sheltered and
unsheltered conditions. Mean daily atmospheric SO2
concentration was 68.7 ^g/m3 and several heavy
rainfalls.
Sample exposed in laboratory to 3 ppm SO2 and 95%
RH at 25 °C for 150 days. Samples were coated with
three carbonaceous particle samples from combustion
sources, and with activated carbon and graphite.
Samples exposed in sheltered ambient environment for
6, 12, or 20 mo.
Samples exposed for 6 mo (cold and hot conditions) in
ambient environment. PM concentrations ranged from
57.3 to 116.7 Mg/m3 (site 1) and 88 to 189.8 ,ug/m3 (site
2). Some exposures also were associated with high SO2,
NO,andNO2.
Samples artificially exposed to fly-ash containing
1309.3 Mg/m3 SO2 (0.5 ppm), at 95% RH and 25 °C for
81 or 140 days. Fly-ash samples from five different
sources were used in study.
Samples exposed to 7,856 ^g/m3 (3 ppm) SO2 at 100%
RH and 25 °C for 30, 60, or 90 days; samples sprayed
with bidistilled water every 7 days to simulate rainfall.
Samples exposed under actual ambient air conditions at
two locations in Rome. Monitoring data obtained for
SO2, NO, NO2, and total suspended particulates (TSP)
but not reported. Exposure was for four seasons.
Significant amounts of gypsum were noted on the Portland Viles (1990)
stone. Sheltered stones also showed soiling by carbonaceous
particles and other combustion products. Etch holes and deep
etching was noted in some of the exposed unsheltered samples.
Exposure to particles from combustion processes enhanced Sabbioni et al.
sulfation of calcareous materials by SO2 because of metal (1996)
content of particles.
Carrara marble found to be more reactive with SO2 than Yerrapragada
Georgia marble possibly because of the compactness of the et al. (1994)
Georgia marble. Greater effects noted when samples were also
exposed to NO2.
Pollutant exposed samples showed increased weight gain over Realini et al.
that expected from natural weathering processes. There was a (1995)
blackening of stone samples exposed to carbonaceous rich
particulate matter.
Exposure to fly-ash did not enhance oxidation of SO2 to Hutchinson
sulfates. Mineral oxides in fly ash contributed to sulphation of et al. (1992)
CaCCv
Exposure to SO2 produced significant quantities of calcium Zappia et al.
sulfite and calcium sulfate on specimens; however, the amount (1994 )
produced was dependent of the porosity, specific surface, and
alkalinity of the sample.
TSP exposure increased the cleaning frequency for stone Lorusso et al.
monuments. Monuments are soiled proportionately overtime, (1997)
based on brightness values. Horizontal surfaces showed higher
graying values because of particle sediment.
-------
TABLE 4-19 (cont'd). CORROSIVE EFFECTS OF PARTICULATE MATTER AND SULFUR DIOXIDE ON STONE
to
o
o
to
Stone
Exposure Conditions
Comments
Source
H
6
o
o
H
O
Limestone
Quartz-cemented
sandstone
Calcite-cemented
sandstone
Granite
Brick
Limestone
Sandstone
Carrara marble
Travertine marble
Irani limestone
Portland limestone
Lime mortar
Pozzolan mortar
Cement mortar
Lime mortar
Carrara marble
Samples from structures exposed for varying periods of
time under ambient air conditions. Samples selected
because of black laer on surface.
Samples of ancient grey crust formed between 1180 and
1636 on the Church of Saint Trophime in Arks and formed
between 1530 and 1187 on the Palazz d'Accursio in
Bolonga.
Samples of the stones and mortars were representative of
those used in the past and currently for new construction
and restorations. Samples were exposed for 6, 12, and
24 mo under ambient conditions in Milan.
Sample of black crust taken from Zamboni Tower Gate.
Samples of crust removed from Milan General Hospital,
built around 1937.
Black layers were found to be primarily comprised of iron
compounds, quartz, silicate, soot, and dirt.
Crust samples contained calcite, soil dust, carbonaceous
particles, and gypsum crystals.
Mortars were more reactive than the stones. Of the mortars,
cement and pozzolan mortar were more reactive than the
lime mortar. Carrara marble was the least reactive of the
stones. The maximum amount of degradation was found in
areas sheltered from rain.
Exposure to environmental pollutants caused the formation
of two separate layers on the mortar: an outer thin surface
black crust composed of gypsum and carbonaceous particles
and the inner composed of products from the dissolution and
sulphation of the carbonate matrix in the mortar.
Gypsum main component of crust followed by carbonaceous
particles and iron oxides. Estimated rate of crust formation
was 2-5 ,um/year. Total amount of gypsum formed over the
lifetime of exposure was 5 to 13 mg/cm2, an estimated
0.2 mg/cm2/year.
Nord and
Ericsson (1993)
Ausset et al.
(1998)
Zappia et al.
(1 998)
Sabbioni et al.
(1998)
Bugini et al.
(2000)
o
HH
H
W
-------
1 of the reaction products at rain runoff sites on the stone. Following sulfur compounds, carbon
2 was reported to be the next highest element in dark crust on historical monuments in Rome.
3 Elemental carbon and organic carbon accounted for 8 and 39% of the total carbon in the black
4 crust samples. The highest percentage of carbon, carbonate carbon, was caused by the carbonate
5 matrix in the stones. The high ratio of organic carbon to elemental carbon indicates the presence
6 of a carbon source other than combustion processes (Ghedini et al., 2000). Cooke and Gibbs
7 (1994) suggested that stones damaged during times of higher ambient pollution exposure likely
8 would continue to exhibit a higher rate of decay, termed the "memory effect," than newer stones
9 exposed under lower pollution conditions. Increased stone damage also has been associated with
10 the presence of sulfur oxidizing bacteria and fungi on stone surfaces (Garcia-Valles et al., 1998;
11 Young, 1996; Saiz-Jimenez, 1993; Diakumaku et al., 1995).
12 Dissolution of gypsum on the stone's surface initiates structural changes in the crust layer.
13 Garica-Valles et al. (1998) proposed a double mechanism: the dissolution of the gypsum, in the
14 presence of sufficient moisture, followed by recrystallization inside fissures or pores. In the
15 event of limited moisture, the gypsum is dissolved and recrystallizes at its original location.
16 According to the authors, this would explain the gypsum-rich crustal materials on stone surfaces
17 sheltered from precipitation.
18 Moisture was found to be the dominant factor in stone deterioration for several sandstones
19 (Petuskey et al., 1995). Dolske (1995) reported that the deteriorative effects of sulfur-containing
20 rain events, sulfates, and SO2 on marble were largely dependent on the shape of the monument or
21 structure rather than the type of marble. The author attributed the increased fluid turbulence over
22 a nonflat vertical surface versus a flat surface to the increased erosion. Sulfur-containing
23 particles also have been reported to enhance the reactivity of Carrara marble and Travertine and
24 Trani stone to SO2 (Sabbioni et al., 1992). Particles with the highest carbon content had the
25 lowest reactivity.
26 The rate of stone deterioration is determined by the pollutant and the pollutant
27 concentration, the stone's permeability and moisture content, and the pollutant deposition
28 velocity. Dry deposition of SO2 between rain events has been reported to be a major causative
29 factor in pollutant-related erosion of calcareous stones (Baedecker et al., 1991; Dolske, 1995;
30 Cooke and Gibbs, 1994; Schuster et al., 1994; Hamilton et al., 1995; Webb et al., 1992). Sulfur
31 dioxide deposition increases with increasing relative humidity (Spiker et al., 1992), but the
April 2002 4-175 DRAFT-DO NOT QUOTE OR CITE
-------
1 pollutant deposition velocity is dependent on the stone type (Wittenburg and Dannecker, 1992),
2 the porosity of the stone, and the presence of hygroscopic contaminants.
3 Although it is clear from the available information that gaseous pollutants, in particular dry
4 deposition of SO2 will promote the decay of some types of stones under the specific conditions,
5 carboneous particles (noncarbonate carbon) may help to promote the decay process by aiding in
6 the transformation of SO2 to a more acidic species (Del Monte and Vittori, 1985). Several
7 authors have reported enhanced sulfation of calcareous material by SO2 in the presence of
8 particles containing metal oxides (Sabbioni et al., 1996; Hutchinson et al., 1992).
9
10 4.4.2 Soiling and Discoloration of Man-Made Surfaces
11 Ambient particles can cause soiling of man-made surfaces. Soiling has been defined as the
12 deposition of particles of less than 10 //m on surfaces by impingement. Soiling generally is
13 considered an optical effect, that is, soiling changes the reflectance from opaque materials and
14 reduces the transmissions of light through transparent materials. Soiling can represent a
15 significant detrimental effect requiring increased frequency of cleaning of glass windows and
16 concrete structures, washing and repainting of structures, and, in some cases, reduction in the
17 useful life of the object. Particles, in particular carbon, also may help catalyze chemical reactions
18 that result in the deterioration of materials during exposure.
19 It is difficult to determine the accumulated particle levels that cause an increase in soiling;
20 however, soiling is dependent on the particle concentration in the ambient environment, particle
21 size distribution, and the deposition rate and the horizontal or vertical orientation and texture of
22 the surface being exposed (Haynie, 1986). The chemical composition and morphology of the
23 particles and the optical properties of the surface being soiled will determine the time at which
24 soiling is perceived (Nazaroff and Cass, 1991). Carey (1959) reported that the average observer
25 could observe a 0.2% surface coverage of black particles on a white background. A recent study
26 suggests that it would take a 12% surface coverage by black particles before there is 100%
27 accuracy in identifying soiling (Bellan et al., 2000). The rate at which an object is soiled
28 increases linearly with time; however, as the soiling level increases, the rate of soiling decreases.
29 The buildup of particles on a horizontal surface is counterbalanced by an equal and opposite
30 depletion process. The depletion process is based on the scouring and washing effect of wind
31 and rain (Schwar, 1998).
April 2002 4-176 DRAFT-DO NOT QUOTE OR CITE
-------
1 4.4.2.1 Stones and Concrete
2 Most of the research evaluating the effects of air pollutants on stone structures have
3 concentrated on gaseous pollutants. The deposition of the sulfur-containing pollutants are
4 associated with the formation of gypsum on the stone (see Section 4.4.1.3). The dark color of
5 gypsum is attributed to soiling by carbonaceous particles from nearby combustion processes.
6 A lighter gray colored crust is attributed to soil dust and metal deposits (Ausset et al., 1998;
7 Camuffo, 1995; Moropoulou et al., 1998). Realini et al. (1995) found the formation of a dark
8 gypsum layer and a loss of luminous reflection in Carrara marble structures exposed for 1 year
9 under ambient air conditions. Dark areas of gypsum were found by McGee and Mossitti (1992)
10 on limestone and marble specimens exposed under ambient air conditions for several years. The
11 black layers of gypsum were located in areas shielded from rainfall. Particles of dirt were
12 concentrated around the edges of the gypsum formations. Lorusso et al. (1997) attributed the
13 need for frequent cleaning and restoration of historic monuments in Rome to exposure to total
14 suspended particulates. They also concluded that, based on a decrease in brightness (graying),
15 surfaces are soiled proportionately over time; however, graying is higher on horizontal surfaces
16 because of sedimented particles. Davidson et al. (2000) evaluated the effects of air pollution
17 exposure on a limestone structure on the University of Pittsburgh campus using estimated
18 average TSP levels in the 1930s and 1940s and actual values for the years 1957 to 1997.
19 Monitored levels of SO2 were available for the years 1980 to 1998. Based on the available data
20 on pollutant levels and photographs, it was thought that soiling began while the structure was
21 under construction. With decreasing levels of pollution, the soiled areas have been slowly
22 washed away, the process taking several decades, leaving a white, eroded surface. Studies
23 describing the effects of particles on stone surfaces are discussed in Table 4-9.
24
25 4.4.2.2 Household and Industrial Paints
26 Few studies are available that evaluate the soiling effects of particles on painted surfaces.
27 Particles composed of elemental carbon, tarry acids, and various other constituents are
28 responsible for soiling of structural painted surfaces. Coarse-mode particles (>2.5 //m) initially
29 contribute more soiling of horizontal and vertical painted surfaces than do fine-mode particles
30 (<2.5 //m), but are more easily removed by rain (Haynie and Lemmons, 1990). The
31 accumulation of fine particles likely promotes remedial action (i.e., cleaning of the painted
April 2002 4-177 DRAFT-DO NOT QUOTE OR CITE
-------
1 surfaces). Coarse-mode particles are primarily responsible for soiling of horizontal surfaces.
2 Rain interacts with coarse particles, dissolving the particle and leaving stains on the painted
3 surface (Creighton et al., 1990; Haynie and Lemmons, 1990). Haynie and Lemmons (1990)
4 proposed empirical predictive equations for changes in surface reflectance of gloss-painted
5 surfaces that were exposed protected and unprotected from rain and oriented horizontally and
6 vertically.
7 Early studies by Parker (1955) and Spence and Haynie (1972) demonstrated an association
8 between particle exposure and increased frequency of cleaning of painted surfaces. Particle
9 exposures also caused physical damage to the painted surface (Parker, 1955). Unsheltered
10 painted surfaces are initially more soiled by particles than sheltered surfaces but the effect is
11 reduced by rain washing. Reflectivity is decreased more rapidly on glossy paint than on flat paint
12 (Haynie and Lemmons, 1990). However, surface chalking of the flat paint was reported during
13 the exposure. The chalking interfered with the reflectance measurements for particle soiling.
14 Particle composition measurements that were taken during exposure of the painted surfaces
15 indicated sulfates to be a large fraction of the fine mode and only a small fraction of the coarse
16 mode. Although no direct measurements were taken, fine mode particles likely also contained
17 large amounts of carbon and possibly nitrogen or hydrogen (Haynie and Lemmons, 1990).
18
19
20 4.5 EFFECTS OF ATMOSPHERIC PARTICIPATE MATTER ON
21 GLOBAL CLIMATE CHANGE PROCESSES AND THEIR
22 POTENTIAL HUMAN HEALTH AND ENVIRONMENTAL IMPACTS
23 Processes causing global climate change and their potential environmental and human
24 health impacts have been accorded extensive attention during the past several decades, and they
25 still continue to be of broad national and international concern. This is reflected by extensive
26 research and assessment efforts undertaken since the mid-1970s by U.S. Federal Government
27 Agencies (e.g., NOAA, EPA, CDC, etc.) or via U.S. Federal Interagency programs (e.g., the U.S.
28 Global Climate Change Research Program [USGCRP]). It is also reflected by analogous
29 extensive research and assessment efforts undertaken by numerous other national governments or
30 international collaborative activities, e.g., those coordinated by the Intergovernmental Panel on
April 2002 4-178 DRAFT-DO NOT QUOTE OR CITE
-------
1 Climate Change (IPCC), established in the 1980s under the joint auspices of the World
2 Meteorological Organization (WMO) and the United Nations Environment Programme (UNEP).
3 Atmospheric particles play important roles in two key types of global climate change
4 processes: (1) alterations in the amount of ultraviolet solar radiation (especially UV-B)
5 penetrating through the Earth's atmosphere and reaching its surface, where it can exert a variety
6 of effects on human health, plant and animal biota, and other environmental components; and
7 (2) alterations in the amount of visible solar radiation transmitted through the Earth's
8 atmosphere. Particles both absorb and reflect solar radiation back into space. The absorption of
9 solar radiation by particles, together with trapping of infrared radiation emitted by the Earth's
10 surface by certain gases, enhances heating of the Earth's surface and lower atmosphere (i.e., the
11 widely-known "greenhouse effect") and leads to consequent "global warming" impacts on human
12 health and the environment. Atmospheric particles also play a lesser role by absorbing infrared
13 radiation emitted by the Earth's surface.
14 The effects of atmospheric PM on the transmission of electromagnetic radiation emitted by
15 the sun at ultraviolet and visible wavelengths and by the earth at infrared wavelengths depend on
16 radiative properties (extinction efficiency, single scattering albedo, and asymmetry parameter) of
17 the particles, which depend, in turn, on the size and shape of the particles, the composition of the
18 particles, and the distribution of components within individual particles. In general, the radiative
19 properties of particles are size- and wavelength-dependent. In addition, the extinction
20 cross-section tends to be at a maximum when the particle radius is similar to the wavelength of
21 the incident radiation. Thus, fine particles present mainly in the accumulation mode would be
22 expected to exert a greater influence on the transmission of electromagnetic radiation than would
23 coarse particles. The composition of particles can be crudely summarized in terms of the broad
24 classes identified in Chapter 2 of this document. These include fine particles consisting mainly
25 of (a) nitrate, sulfate, mineral dust, elemental carbon, organic carbon compounds (e.g., PAHs),
26 and (b) metals derived from high temperature combustion or smelting processes. The major
27 sources of these components are shown in Table 3-9 of Chapter 3 in this document.
28 Knowledge of factors controlling the transfer of solar radiation in the ultraviolet spectral
29 range is needed for assessing potential biological and environmental impacts associated with
30 exposure to UV-B radiation (290 to 315 nm). Knowledge of the effects of PM on the transfer of
31 radiation in the visible and infrared spectral regions is needed for assessing the relationship
April 2002 4-179 DRAFT-DO NOT QUOTE OR CITE
-------
1 between particles and global warming and its environmental and biological impacts. Important
2 conceptual aspects and factors related to solar ultraviolet radiation processes and effects are first
3 summarized below and atmospheric PM roles discussed. This is followed by a summary of
4 global warming processes, their potential human health and environmental impacts, and their
5 potential relationships to atmospheric PM.
6
7 4.5.1 Solar Ultraviolet Radiation Transmission Impacts on Human Health
8 and the Environment: Atmospheric Particulate Matter Effects
9 4.5.1.1 Potential Effects of Increased Ultraviolet Radiation Transmission
10 The transmission of solar UV-B radiation through the earth's atmosphere is controlled by
11 ozone, clouds, and particles. The depletion of stratospheric ozone caused by the release of
12 anthropogenically produced chlorine (Cl)-and bromine (Br)-containing compounds has resulted
13 in heightened concern about potentially serious increases in the amount of solar UV-B radiation
14 (SUVB) reaching the Earth's surface. SUVB is also responsible for initiating the production of
15 OH radicals that oxidize a wide variety of volatile organic compounds, some of which can
16 deplete stratospheric ozone (e.g., CH3C1, CH3Br), absorb terrestrial infrared radiation (e.g., CH4),
17 and contribute to photochemical smog formation (e.g., C2H4, C5H8).
18 Increased penetration of SUVB to the Earth's surface as the result of stratospheric ozone
19 depletion continues to be of much concern because of proj ections of consequent increased
20 surface-level SUVB exposure and associated potential negative impacts on human health, plant
21 and animal biota, and man-made materials. Several summary overviews (Kripke, 1989; Grant,
22 1989; Kodama and Lee, 1994; Van der Leun et al., 1995, 1998) of salient points related to
23 stratospheric ozone depletion and bases for concern provide a concise introduction to the subject,
24 as does Figure 4-31. As shown to the left in Figure 4-31, stratophospheric ozone depletion
25 results from: (a) anthropogenic emissions of certain trace gases having long atmospheric
26 residence times, e.g., chlorofluorocarbons (CFCs), carbon tetrachloride (CC14), and Halon 1211
27 (CF2C1 Br) and 1301 (CF3Br)—which have atmospheric residence times of 75 to 100 years,
28 50 years, 25 years, and 110 years, respectively; (b) their tropospheric accumulation and gradual
29 transport, over decades, up to the stratosphere, where (c) they photolyze to release Cl and Br that
30 catalyze ozone destruction; leading to (d) stratospheric ozone depletion. Such ozone depletion is
31 most marked over Antarctica during spring in the Southern Hemisphere, to a less marked but still
April 2002 4-180 DRAFT-DO NOT QUOTE OR CITE
-------
BASES FOR CONCERN ABOUT STRATOSPHERIC OZONE DEPLETION
DUE TO CFC's, HALONS, AND OTHER TRACE GASES
Stratospheric
Ozone Depletion
Cl, Br Catalyze
Ozone Reduction
Photodissociation
Releases Cl and Br
Slow Transport
to Stratosphere
Tropospheric
Accumulation
Air Emissions of
CFC's, Halons, etc.
OZONE DEPLETION EFFECTS
/CFC's &O3 Column^
I Reorganization I
Climate Changes:
Temp., Winds, etc.
Increase of Air
Stagnation Periods
Accumulation of
Tropospheric O3
and Acid Aerosols
Increased UV-B Radiation
Penetration to Surface
Altered Bio-
geochemical
Cycling
UV-B Radiation Direct
Human Health Impacts
Immune
Suppresssion
Natural Ecosystem
and Agriculture
Impacts
Skin Damage
(Sunburn)
Damage
to Eye
Terrestrial
Ecosystem Shifts
Lower Crop Yields
AND
Man's Production
of CFC's, Halons,
Other Trace Gasses
Environmental
Effects: Crop,
Forest Damage
Human
Health
Effects
Infectious
Diseases
Increased
Skin Cancer
Premature
Skin Aging
Cataracts
Incidence
Increased
Aquatic
Ecosystem Shifts
Less Plankton &
Seafood
Figure 4-31. Processes involved in stratospheric ozone depletion because of man's
production of CFCs, halons, and other trace gases are shown to the left. The
types of effects caused by stratospheric ozone depletion and consequent
increased UV-B penetration to the Earth's surface are hypothesized to include
both direct effects on human health (e.g., increased cancer rates, immune
suppression, etc.) and other terrestrial and aquatic ecological effects resulting
from increased UV-B alterations of biogeochemical cycles.
1 significant extent over the Arctic Polar Region during late winter and spring in the Northern
2 Hemisphere, and to a lesser extent, over mid-latitude regions during any season.
3 Given the long time involved in transport of such gases to the stratosphere and their long
4 residence times there, any effects already seen on stratospheric ozone are likely caused by the
5 atmospheric loadings of trace gases from anthropogenic emissions over the past few decades.
April 2002
4-181
DRAFT-DO NOT QUOTE OR CITE
-------
1 Those gases already in the atmosphere may continue to exert stratospheric ozone depletion
2 effects well into the 21st century. Shorter-lived gases, such as CH3Br, also exert significant
3 ozone depletion effects.
4 The main types of deleterious effects hypothesized as likely to result from stratospheric
5 ozone depletion and consequent increased SUVB penetration through the Earth's atmosphere
6 include the following.
7 (1) Direct Human Health Effects, such as skin damage (sunburn), leading to more rapid aging
8 and increased incidence of skin cancer; ocular effects (retinal damage and increased cataract
9 formation possibly leading to blindness); and suppression of some immune system
10 components (contributing to skin cancer induction and spread to nonirradiated skin areas,
11 as well as possibly increasing susceptibility to certain infectious diseases or decreasing
12 effectiveness of vaccinations).
13 (2) Agricultural/Ecological Effects, mediated largely through altered biogeochemical cycling
14 resulting in consequent damaging impacts on terrestrial plants (leading to possible reduced
15 yields of rice, other food crops, and commercially important trees, as well as to biodiversity
16 shifts in natural terrestrial ecosystems); and deleterious effects on aquatic life (including
17 reduced ocean zooplankton and phytoplankton, as important base components of marine
18 food-chains supporting the existence of commercially important, edible fish and other
19 seafood, as well as to other aquatic ecosystem shifts).
20 (3) Indirect Human Health and Ecological Effects, mediated through increased tropospheric
21 ozone formation (and consequent exacerbation of surface-level, ozone-related health and
22 ecological impacts) and alterations in the concentrations of other important trace species,
23 most notably the hydroxyl radical and acidic aerosols.
24 (4) Other Types of Effects, such as faster rates of polymer weathering because of increased
25 UV-B radiation and other effects on man-made commercial materials and cultural artifacts,
26 secondary to climate change or exacerbation of air pollution problems.
27 Extensive qualitative and quantitative characterizations of stratospheric ozone depletion
28 processes and projections of their likely potential impacts on human health and the environment
29 have been the subjects of periodic (1988, 1989, 1991, 1994, 1998) international assessments
30 carried out under WMO and UNEP auspices since the 1987 signing of the Montreal Protocol on
31 Substances that Deplete the Ozone Layer. For detailed up-to-date information, the reader is
April 2002 4-182 DRAFT-DO NOT QUOTE OR CITE
-------
1 referred to recent international assessments of (a) processes contributing to stratospheric ozone
2 depletion and the status of progress towards ameliorating the problem (WMO, 1999) and
3 (b) revised qualitative and quantitative projections of likely consequent human health and
4 environmental effects (UNEP, 1998, 2000) — with the findings and conclusions of these
5 assessments being incorporated herein by reference.
6 Of considerable importance is the growing recognition, as reflected in these newer
7 assessments, of impacts of enhanced solar radiation on biogeochemical cycles (see, for example,
8 Zepp et al., 1998, and earlier discussions in this chapter in Section 4.2). As noted in the Zepp
9 et al. paper, the effects of UV-B radiation (both in magnitude and direction) on trace gas (e.g.,
10 CO) emissions and mineral nutrient cycling are species specific and can affect a variety of
11 processes. These include, for example, changes in the chemical composition of living plant
12 tissue, photodegradation of dead plant matter (e.g., ground litter), release of CO from vegetation
13 previously charred by fire, changes in microbial decomposer communities, and effects on
14 nitrogen-fixing microorganisms and plants. Also, studies of natural acquatic ecosystems indicate
15 that organic matter is the primary determinant of UV-B penetration through water. Changes in
16 the amount and composition of organic matter, caused by enhanced UV-B penetration, affect the
17 transmission of solar ultraviolet and visible radiation through the water column. These changes
18 in light quality broadly impact the effects of UV-B on aquatic biogeochemical cycles. Enhanced
19 UV-B levels have both positive and negative impacts on microbial activities in aquatic
20 ecosystems that can affect nutrient cycling and the uptake or release of greenhouse gases. Thus,
21 there are emerging complex issues regarding interactions and feedbacks between climate change
22 and changes in terrestrial and marine biogeochemical cycles because of increased UV-B
23 penetration to the Earth's surface.
24 In contrast to the above types of negative impacts projected as likely to be associated with
25 increased UV-B penetration to Earth's surface, some research results are suggestive of possible
26 beneficial effects of increased UV-B radiation. For example, a number of U.S. and international
27 studies have focused on the protective effects of UV-B radiation with regard to non-skin cancer
28 incidence. One of the first of these studies investigated potential relationships between sunlight,
29 vitamin D and colon cancer (Garland and Garland, 1980). More recent studies continue to
30 provide evidence that UV-B radiation may be protective against several types of cancer and some
31 other diseases. For example, Grant (2002) has conducted a number of ecologic-type
April 2002 4-183 DRAFT-DO NOT QUOTE OR CITE
-------
1 epidemiologic studies, which suggest that UV-B radiation, acting through the production of
2 vitamin D, is a risk-reduction factor for mortality due to several types of cancer, including cancer
3 of the breast, colon, ovary, and prostate, as well as non-Hodgkin lymphoma. Other related
4 studies that provide evidence for protective effects of UV-B radiation include: Gorham et al.
5 (1989); Gorham et al. (1990); Garland et al. (1990); Hanchette and Schwartz (1992); Ainsleigh
6 (1993); Lefkowitz and Garland (1994); Hartge et al. (1996); and Freedman et al. (1997).
7 As noted in the above detailed international assessments, since the signing of the Montreal
8 Protocol, much progress has been made in reducing emissions of ozone depleting gases, leading
9 to estimates that the maximum extent of stratospheric ozone depletion has likely leveled off
10 during recent years, and this is expected to be followed by gradual lessening of the problem and
11 its impacts during the next half-century. However, the assessments also note that the modeled
12 projections are subject to considerable uncertainty (see, for example, UNEP, 2000). Varying
13 potential roles of atmospheric particles, discussed below, are among numerous salient factors
14 complicating predictive modeling efforts.
15
16 4.5.1.2 Effects of Airborne Particles on Transmission of Solar Ultraviolet Radiation
17 Through the Atmosphere
18 A given amount of ozone in the lower troposphere has been shown to absorb more solar
19 radiation than an equal amount of ozone in the stratosphere because of the increase in its
20 effective optical path produced by Rayleigh scattering in the lower atmosphere (Briihl and
21 Crutzen, 1988). The effects of particles are more complex. The impact of particles on the SUVB
22 flux throughout the boundary layer are highly sensitive to the altitude of the particles and to their
23 single scattering albedo. Even the sign of the effect can reverse as the composition of the particle
24 mix changes from scattering to absorbing types (e.g., from sulfate to elemental carbon or PAHs)
25 (Dickerson et al., 1997). In addition, scattering by particles also may increase the effective
26 optical path of absorbing molecules, such as ozone, in the lower atmosphere.
27 The effects of particles present in the lower troposphere on the transmission of SUVB have
28 been examined both by field measurements and by radiative transfer model calculations. The
29 presence of particles in urban areas modifies the spectral distribution of solar irradiance at the
30 surface. Shorter wavelength radiation (i.e., in the ultraviolet) is attenuated more than visible
31 radiation (e.g., Peterson et al., 1978; Jacobson, 1999). Wenny et al. (1998) also found greater
April 2002 4-184 DRAFT-DO NOT QUOTE OR CITE
-------
1 attenuation of SUVB than SUVA (315 to 400 nm). However, this effect depends on the nature
2 of the specific particles involved and, therefore, is expected to depend strongly on location.
3 Lorente et al. (1994) observed an attenuation of SUVB ranging from 14 to 37%, for solar zenith
4 angles ranging from about 30° to about 60°, in the total (direct and diffuse) SUVB reaching the
5 surface in Barcelona during cloudless conditions on very polluted days (aerosol scattering optical
6 depth at 500 nm, 0.46 < T500nm < 1.15) compared to days on which the turbidity of urban air was
7 similar to that for rural air (T500nm ^ 0.23). Particle concentrations that can account for these
8 observations can be estimated roughly by combining Koschmeider's relation for expressing
9 visual range in terms of extinction coefficient with one for expressing the mass of PM2 5 particles
10 in terms of visual range (Stevens et al., 1984). By assuming a scale height (i.e., the height at
11 which the concentration of a substance falls off to 1/e of its value at the surface) of 1 km for
12 PM2 5, an upper limit of 30 //g/ m3 can be derived for the clear case and between 60 and
13 150 //g/m3 for the polluted case. Estupifian et al. (1996) found that summertime haze under clear
14 sky conditions attenuates SUVB between 5 and 23% for a solar zenith angle of 34°, compared to
15 a clear sky day in autumn. Minis (1996) measured a decrease in SUVB by about 80% downwind
16 of major biomass burning areas in Amazonia in 1995. This decrease in transmission
17 corresponded to optical depths at 340 nm ranging from three to four. Justus and Murphey (1994)
18 found that SUVB reaching the surface decreased by about 10% because of changes in aerosol
19 loading in Atlanta, GA, from 1980 to 1984. Also, higher particle levels in Germany (48 °N) may
20 be responsible for greater attenuation of SUVB than in New Zealand (Seckmeyer and McKenzie,
21 1992).
22 In a study of the effects of nonurban haze on SUVB transmission, Wenny et al. (1998)
23 derived a very simple regression relation between the measured aerosol optical depth at 312 nm,
24
25 ln( SUVB transmission at solar noon) = -0.1422 T312nm- 0.138, R2 = 0.90, (4-11)
26
27 and the transmission of SUVB to the surface. In principle, values of T312nm could be found from
28 knowledge of the aerosol optical properties and visual range values. Wenny et al. (1998) also
29 found that absorption by particles accounted for 7 to 25% of the total (scattering + absorption)
30 extinction. Relations such as the above one are strongly dependent on local conditions and
31 should not be used in other areas without knowledge of the differences in aerosol properties.
April 2002 4-185 DRAFT-DO NOT QUOTE OR CITE
-------
1 Although all of the above studies reinforce the idea that particles play a major role in modulating
2 the attenuation of SUVB, none included measurements of ambient PM concentrations, so direct
3 relations between PM levels and SUVB transmission could not be determined.
4 Vuilleumier et al. (2001) concluded that variations in aerosol scattering and absorption were
5 responsible for 97% of the variability in the optical depth measured at seven wavelengths from
6 300 to 360 nm at Riverside, CA from 1 July to 1 November, 1997. Similar measurements made
7 at Mt. Wilson, located above the main surface haze layer, showed that 80% of the variations in
8 optical depth were still driven by variations in aerosol scattering and absorption. The remainder
9 of the variability in optical depth was attributed mainly to variability in ozone under clear-sky
10 conditions. However, these results cannot be extrapolated to other locations because these
11 effects are coupled and non-linear and are not straightforward. They depend on the
12 concentrations of these species and on the physical and chemical characteristics of the particles.
13 Hence, any quantitative statements regarding the relative importance of particles and ozone will
14 be location-specific.
15 Liu et al. (1991) estimated, roughly, the overall effects on atmospheric transmission of
16 SUVB of increases of anthropogenic airborne particles that have occurred since the beginning of
17 the industrial revolution. Based on (a) estimates of the reduction in visibility from about 95 km
18 to about 20 km over nonurban areas in the eastern United States and in Europe, (b) calculations
19 of optical properties of airborne particles found in rural areas to extrapolate the increase in
20 extinction at 550 to 310 nm, and (c) radiative transfer model calculations, Liu et al. concluded
21 that the amount of SUVB reaching Earth's the surface likely has decreased from 5 to 18% since
22 the beginning of the industrial revolution. This was attributed mainly to scattering of SUVB
23 back to space by sulfate containing particles. Radiative transfer model calculations have not
24 been done for urban particles.
25 Although aerosols are expected to decrease the flux of SUVB reaching the surface,
26 scattering by particles is expected to result in an increase in the actinic flux within and above the
27 aerosol layer. However, when the particles significantly absorb SUVB, a decrease in the actinic
28 flux is expected. Actinic flux is the radiant energy integrated over all directions at a given
29 wavelength incident on a point in the atmosphere, and is the quantity needed to calculate rates of
30 photolytic reactions in the atmosphere. Blackburn et al. (1992) measured attenuation of the
31 photolysis rate of ozone and found that aerosol optical depths near unity at 500 nm reduced
April 2002 4-186 DRAFT-DO NOT QUOTE OR CITE
-------
1 ozone photolysis rate by as much as a factor of two. Dickerson et al. (1997) showed that the
2 photolysis rate for NO2, a key parameter for calculating the overall intensity of photochemical
3 activity, could be increased within and above a scattering aerosol layer extending from the
4 surface, although it would be decreased at the surface. This effect is qualitatively similar to what
5 is seen in clouds, where photolysis rates are increased in the upper layers of a cloud and above
6 the cloud (Madronich, 1987). For a simulation of an ozone episode that occurred during July
7 1995 in the Mid-Atlantic region, Dickerson et al. (1997) calculated ozone increases of up to
8 20 ppb compared to cases that did not include the radiative effects of particles in urban airshed
9 model (UAM-IV) simulations. In contrast, Jacobson (1998) found that particles may have
10 caused a 5 to 8% decrease in O3 levels during the Southern California Air Quality Study in 1987.
11 Absorption by organic compounds and nitrated inorganic compounds was hypothesized to
12 account for the reductions in UV radiation intensity.
13 The photolysis of ozone in the Hartley bands also leads to production of electronically
14 excited oxygen atoms, O(JD) that then react with water vapor to form OH radicals. Thus,
15 enhanced photochemical production of ozone is accompanied by the scavenging of species
16 involved in greenhouse warming and stratospheric depletion. However, these effects may be
17 neutralized or even reversed by the presence of absorbing material in the particles. Any
18 evaluation of the effects of particles on photochemical activity therefore will depend on the
19 composition of the particles and also will be location-specific.
20 Also complicating any straightforward evaluation of UV-B penetration to specific areas of
21 the Earth's surface are the influences of clouds, as discussed by Erlick et al. (1998), Frederick
22 et al. (1998), and Soulen and Fredrick (1999). The transmission of solar UV and visible
23 radiation is highly sensitive to cloud type and cloud amount and the extent of their external or
24 internal mixing with cloud droplets. Even in situations of very low atmospheric PM (e.g., over
25 Antarctica), interannual variations in cloudiness over specific areas can be as important as ozone
26 levels in determining UV surface irradiation, with net impacts varying from a month or season to
27 another (Soulen and Fredrick, 1999). Evaluations of the effects of changes in the transmission of
28 solar UV-B radiation to the surface have been performed usually for cloud-free or constant
29 cloudiness conditions.
30 Given the above considerations, quantification of projected effects of variations in
31 atmospheric PM on human health or the environment because of the effects of particles on the
April 2002 4-187 DRAFT-DO NOT QUOTE OR CITE
-------
1 transmission of solar UV-B radiation requires location-specific evaluations, taking into account
2 composition, concentration, and internal structure of the particles; temporal variations in
3 atmospheric mixing height and depths of layers containing the particles the abundance of ozone
4 and other absorbers within the planetary boundary layer and the free troposphere. The outcome
5 of such modeling effects would likely vary from location to location in terms of increased or
6 decreased surface level UV-B exposures because of location-specific changes in atmospheric PM
7 concentrations or composition. For example, to the extent that any location-specific scattering by
8 airborne PM were to affect the directional characteristics of UV radiation at ground level, and
9 thereby enhance radiation incident from low angles (Dickerson, 1997), the biological
10 effectiveness (whether deleterious or beneficial) of resulting ground-level UV-B exposures could
11 be enhanced. Airborne PM also can reduce the ground-level ratio of photorepairing radiation
12 (UV-A and short-wavelength visible) to damaging UV-B radiation. Lastly, PM deposition is a
13 major source of PAHs in certain freshwater lakes and coastal areas, and the adverse effects of
14 solar UV are enhanced by the uptake of PAHs by aquatic organisms. Thus, although airborne
15 PM may, in general, tend to reduce ground-level UV-B, its net effect in some locations may be to
16 increase UV damage to certain aquatic and terrestrial organisms, as discussed by Cullen and
17 Neale (1997).
18
19 4.5.2 Global Warming Processes, Human Health and Environmental
20 Impacts, and Roles of Atmospheric Particle
21 4.5.2.1 Bases for Concern Regarding Global Warming and Climate Change
22 Various trace gases emitted by man's activities, including several noted above as
23 contributing to stratospheric ozone depletion, can act as "greenhouse gases" (GHG). That is, as
24 their tropospheric concentrations increase, they retard the escape of infrared radiation from the
25 earth's surface and thereby contribute to the trapping of heat near the surface (the "greenhouse
26 effect") and, ultimately, to consequent global warming and climate change. Much concern has
27 evolved with regard to increases in the naturally very low concentrations in the atmosphere of
28 some of these gases, especially carbon dioxide (CO2), nitrous oxide (N2O), methane (CH4),
29 chloroflurocarbons (CFCs), and tropospheric ozone (O3).
30 Atmospheric processes involved in mediating global warming and its likely consequent
31 effects have been reviewed extensively (United Nations Environment Programme, 1986; World
April 2002 4-188 DRAFT-DO NOT QUOTE OR CITE
-------
1 Meteorological Organization, 1988; U.S. Environmental Protection Agency, 1987; IPCC, 1996,
2 1998, 2001a,b; NAST, 2000) and more concisely summarized by others (e.g., Patz et al.,
3 2000a,b). The reader is referred to such reviews for more detailed information than that
4 concisely summarized below. The main focus here is first (a) to provide a brief summary of key
5 points regarding processes involved and types of effects projected as likely to be associated with
6 global warming and climate change and, then, (b) to discuss salient considerations regarding
7 potential impacts of atmospheric PM on such processes and effects.
8 The Third Assessment Report of the IPCC (200 la) discusses observed past changes in the
9 climate system of the Earth. Of particular note is the calculation stated in that IPCC report
10 indicating that the global average temperature (i.e., the average of near surface air temperatures
11 over land and sea surface temperatures) has increased by 0.6 ±0.2 °C over the course of the
12 20th Century. Globally, the decade of the 1990's was likely the warmest of the Century and
13 1998 the warmest year in the instrumental record (IPCC, 200la). New analyses of proxy data for
14 the Northern Hemisphere also indicate that the rise in temperature over the 20th Century is likely
15 the largest of any century during the past 1,000 years (IPCC, 200la). However, the projecting of
16 future trends in global average temperature and regional climate impacts is difficult and fraught
17 with many uncertainties.
18 All of the above noted assessments and summaries emphasize that estimating likely future
19 global warming trends and associated climate change caused by greenhouse gases is extremely
20 complex, with modeling results being highly dependent on key assumptions about the rates of
21 future increases in various gases and numerous other factors (including particle effects).
22 Modeling of the magnitude of the warming directly associated with radiative forcing by
23 greenhouse gases (without feedback enhancement) projects temperature increases ranging from
24 1.4 °C to 5.8 °C over the period 1990 to 2100 (IPCC, 200la). This range does not include
25 uncertainties in the modeling of radiative forcing (e.g., aerosol radiative forcing). Feedbacks that
26 likely would increase temperatures further are expected to occur. Increased water vapor
27 (trapping heat) and snow and ice melting (reducing reflection of radiation back into space) are
28 two examples of such feedback factors expected to increase temperatures. However, major
29 uncertainties exist with regard to feedbacks between global warming and clouds, which could
30 either amplify or, perhaps, reduce a temperature rise. Taking assumptions about rates of increase
31 (or decrease) in GHG concentrations, consequent initial warming effects, feedback effects, and
April 2002 4-189 DRAFT-DO NOT QUOTE OR CITE
-------
1 accompanying uncertainties into account, numerous modeling efforts have attempted to project
2 likely future trends in global warming. Despite the complexity and uncertainties inherent in such
3 modeling efforts, all typically agree that some global warming has occurred and will continue to
4 occur during the coming decades, but the ranges of quantitative estimates vary considerably
5 depending on specific assumptions incorporated into the models. Thus, for example, "low"
6 scenarios assuming stabilization or reductions in GHG emissions (resulting from implementation
7 of the 1987 Montreal Protocol) project lower temperature changes than other scenarios assuming
8 higher rates of increase in GHG emissions or differing feedback-effect patterns.
9 Given the wide range of estimates of global warming trends and patterns of associated
10 climate change emerging from modeling efforts, the estimation of likely human health and
11 ecological effects associated with global warming on any quantitative basis is extremely difficult.
12 The onset of any notable global warming effect is also important, with various analyses
13 indicating that global temperatures for the past century have been rising (and now appear to be
14 beyond average levels within the range of variation seen with cycles of global warming or
15 cooling over the past several centuries before marked anthropogenic emissions of greenhouse
16 gases occurred). Also posing difficulties for the quantitative estimation of human health and
17 other effects are expected wide regional variations in temperature and climate characteristics
18 (e.g., rain and snowfall amounts) that may be projected reasonably to result from various global
19 warming trend scenarios. Lastly, it should be noted that, despite general warming trends in
20 long-term average temperatures, wide extremes in both high and low temperatures also are
21 expected to occur more frequently in some areas.
22 Special reports of the IPCC Working Group II on impacts of climate change (IPCC, 1998,
23 200 Ib) assess global warming processes and identify several types of vulnerabilities likely to
24 occur because of climate change resulting from global warming. Such general types of
25 vulnerabilities include impacts on terrestrial and aquatic ecosystems, hydrology and water
26 resources, food and fiber production, coastal systems, and human health. The executive
27 summaries of these IPCC (1998, 2001b) reports provide helpful overviews of key points
28 regarding projected global warming processes, likely climate change patterns, and their
29 consequent impacts in terms of the types of vulnerabilities noted above.
30 The IPCC (1998, 2001b) reports indicate that human activities resulting in emissions of
31 long-lived GHCs are projected by general circulation models (GCMs) as likely to lead to marked
April 2002 4-190 DRAFT-DO NOT QUOTE OR CITE
-------
1 global and regional changes in temperature, precipitation and other climate variables. The
2 average rate of warming is projected to be more rapid than any seen in the past 10,000 years,
3 although regional changes could differ substantially from mean global rates. This is expected to
4 result in increases in global mean sea level; prospects for more extreme weather events, floods,
5 and droughts in some areas; and consequent changes in soil moisture. The most recent IPCC
6 Reports (2001a,b) highlight GCM modeling results, based on various scenarios of current and
7 plausible future emissions of GHGs and aerosols and the range of sensitivities of climate change
8 to atmospheric levels (and residence time) of GHGs, which project mean annual global surface
9 temperature increases leading by 2100 to global mean sea level rise of 0.09 to 0.88 m above 1990
10 levels and significant changes in spatial and temporal patterns of precipitation.
11 Human health, ecosystems, and socioeconomic sectors (e.g., hydrology and water resources,
12 food and fiber production, etc.) are also projected by IPCC (1998, 2001b) to be vulnerable to the
13 magnitude and rate of climate change and extremes (see Table 4-20). Wide variations in the
14 courses and net impacts of climate change in different geographic areas can be expected.
15 In general, projected climate change impacts can be expected to represent additional stresses on
16 those natural ecosystems and human societal systems already impacted by increasing resource
17 demands, unsustainable resource management practices, and pollution—with wide variation
18 likely across regions and nations in their ability to cope with consequent alterations in ecological
19 balances, in availability of adequate food, water, and clean air, and in human health and safety.
20 However, although many regions are likely to experience severe adverse impacts (some possibly
21 irreversible) of climate change, some climate change impacts may be locally beneficial in some
22 regions.
23 The 1998 IPCC special report regarding the assessment of different types of vulnerabilities
24 to climate change included projections of likely impacts for each of 10 different geographic
25 regions of the Earth, including those projected for two regions (North America and Polar) of
26 most relevance to the continental United States and Alaska. Probably of most note are findings
27 indicating that (a) the characteristics of subregions and sectors of North America suggest that
28 neither impacts of climate change nor response options will be uniform, and (b) many systems of
29 North America are moderately to highly sensitive to climate change, with the range of estimated
30 effects encompassing potential substantial damage or, conversely, some potential for beneficial
31 outcomes. The most vulnerable continental United States sectors and regions include long-lived
April 2002 4-191 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 4-20. EXAMPLES OF IMPACTS RESULTING FROM PROJECTED
CHANGES IN EXTREME CLIMATE EVENTS
Projected changes during the 21st Century in Extreme
Climate Phenomena and their Likelihood3
Representative Examples of Projected Impacts'1
(all high confidence of occurrence in some areas')
Simple Extremes
Higher maximum temperatures; more hot days and heat
waves4 over nearly all land areas (very likely)
Increased incidence of death and serious illness in older age
groups and urban poor
Increased heat stress in livestock and wildlife
Shift in tourist destinations
Increased risk of damage to a number of crops
Increased electric cooling demand and reduced energy supply
reliability
Higher (increasing) minimum temperatures; fewer cold
days, frost days, and cold waves4 over nearly all land
areas (very likely?)
Decreased cold-related human morbidity and mortality
Decreased risk of damage to a number of crops, and increased
risk to others
Extended range and activity of some pest and disease vectors
Reduced heating energy demand
More intense precipitation events (very likely" over
many years)
Increased flood, landslide, avalance, and mudslide damage
Increased soil erosion
Increased flood runoff could increase recharge of some
floodplain aquifers
Increased pressure on government and private flood insurance
systems and disaster relief
Complex Extremes
Increased summer drying over most mid-latitude
continental interiors and associated risk of drought
(likely*)
Decreased crop yields
Increased damage to building foundations caused by ground
shrinkage
Decreased water resource quantity and quality
Increased risk of forest fire
Increase in tropical cyclone peak wind intensities, mean
and peak precipitation intensities (likely" over some
areas)6
Increased risk to human life, risk of infections, disease
epidemics, and many other risks
Increased coastal erosion and damage to coastal buildings and
infrastructure
Increased damage to coastal ecosystems such as coral reefs and
mangroves
Intensified droughts and floods associated with El Nino
events in many different regions (likely") (see also under
droughts and intense precipitation events)
Decreased agricultural and rangeland productivity in drought-
and flood-prone regions
Decreased hydro-power potential in drought-prone regions
Increased Asian summer monsoon precipitation
variability (likely")
Increased flood and drought magnitude and damages in
temperate and tropical Asia
Increased intensity of mid-latitude storms
(little agreement between current models)4
Increased risks to human life and health
Increased property and infrastructure losses
Increased damage to coastal ecosystems
^Likelihood refers to judgmental estimates of confidence used by TAR WGI: very likely (90-99% chance); likely (66-90%
chance). Unless otherwise stated, information on climate phenomena is taken from the Summary for Policymakers,
TAR WGI. TAR WGI = Third Assessment Report of Working Group 1 (IPCC, 200 la).
bThese impacts can be lessened by appropriate response measures.
cHigh confidence refers to probabilities between 67 and 95%.
Information from TAR WGI, Technical Summary.
'Changes in regional distribution of tropical cyclones are possible but have not been established.
Source: IPCC (2001 b).
April 2002
4-192
DRAFT-DO NOT QUOTE OR CITE
-------
1 natural forest ecosystems in the East and interior West, water resources in the southern plains,
2 agriculture in the Southeast and southern plains, northern ecosystems and habitats, estuaries and
3 beaches in developed areas, and low-latitude cool and cold water fisheries. Other sectors or
4 subregions may benefit from warmer temperatures or increased CO2 fertilization (e.g., west coast
5 coniferous forests; some western rangelands; reduced energy costs for heating in northern
6 latitudes; reduced road salting and snow-clearance costs; longer open-water seasons in norther
7 channels and ports; and agriculture in the northern latitudes, the interior West, and the west
8 coast). For Alaska, substantial shifts in ecosystems (with possible major declines or loss of some
9 sensitive species like bear and caribou or of other ice-dependent animals) may occur in parallel to
10 beneficial effects such as opening of ice-bound water transportation routes or possible expanded
11 agricultural viability secondary to longer growing seasons. On the other hand, for North
12 America, the potential for mainly deleterious direct or indirect effects on human health is likely
13 to increase (e.g., increased mortality directly linked to temperature extremes, increases in
14 incidence and spread of vector-borne infectious diseases, impacts secondary to sea-level rise, and
15 impacts secondary to increased tropospheric air pollution.
16 More detailed evaluations of possible global climate change impacts on various U.S.
17 geographic areas are being conducted by the United States Global Change Research Program
18 (USGCRP). An overview report on the assessment results and key findings from a series of
19 workshops convened by the USGCRP National Assessment Synthesis team (NAST) has been
20 prepared (NAST, 2000). Overall key findings from the USGCRP (NAST, 2000) report are noted
21 below.
22 (1) Increased Warming. Assuming continued growth in world GHG emissions, the primary
23 climate models used in the USGCRP assessment project that temperatures in the United
24 States will rise by 5 to 9 °F (3 to 5 °C) on average during the next 100 years. A wide
25 range of outcomes is possible.
26 (2) Differ ing Regional Impacts. Climate change will vary widely across the United States.
27 Temperature increases will vary somewhat from region to region. Heavy and extreme
28 precipitation events are likely to become more frequent, yet some regions will get drier.
29 The potential impacts of climate change will vary widely across the nation.
30 (3) Vulnerable Ecosystems. Many ecosystems are highly vulnerable to the projected rate and
31 magnitude of climate change. A few, such as alpine meadows in the Rocky Mountains and
April 2002 4-193 DRAFT-DO NOT QUOTE OR CITE
-------
1 some barrier islands, are likely to disappear entirely in some areas; and others, such as
2 some forests of the Southeast, are likely to experience major species shifts or break up into
3 a mosaic of grasslands, woodlands, and forests. Goods and services lost through
4 disappearance or fragmentation of certain ecosystems are likely to be costly or impossible
5 to replace.
6 (4) Widespread Water Concerns. Water is an issue in every region, but the nature of the
7 vulnerabilities varies, with different nuances in each. Drought is an important concern in
8 every region. Floods and water quality are concerns in many regions. Snowpack changes
9 are especially important in the West, the Pacific Northwest, and Alaska.
10 (5) Secure Food Supply. At the national level, the U.S. agriculture sector is likely to be able to
11 adapt to climate change. Overall, U.S. crop productivity is very likely to increase over the
12 next few decades, but the gains will not be uniform across the nation. Falling prices and
13 competitive pressures are very likely to stress some farmers, while benefiting consumers.
14 (6) Near-Term Increases in Forest Growth. Forest productivity is likely to increase over the
15 next several decades in some areas as trees respond to higher CO2 levels. Over the longer
16 term, changes in larger scale processes such as fire, insects, droughts, and disease will
17 possibly decrease forest productivity. Also, climate change is likely to cause long-term
18 shifts in forest species (e.g., distribution of sugar maple stands more northward, out of the
19 United States).
20 (7) Increased Damage in Coastal and Permafrost Areas. Climate change and the resulting
21 rise in sea level are likely to exacerbate threats to building, roads, powerlines, and other
22 infrastructure in climatically sensitive places. For example, infrastructure damage is
23 related to permafrost melting in Alaska and to sea level rise and storm surges in low-lying
24 coastal areas.
25 (8) Adaptation Determines Health Outcomes. A range of negative health impacts is possible
26 from climate change, but adaptation is likely to help protect much of the U.S. population.
27 Maintaining our nation's public health and community infrastructure, from water treatment
28 systems to emergency shelters, will be important for minimizing the impacts of waterborne
29 diseases, heat stress, air pollution, extreme weather events, and diseases transmitted by
30 insects, ticks, and rodents.
April 2002 4-194 DRAFT-DO NOT QUOTE OR CITE
-------
1 (9) Other Stresses Magnified by Climate Change. Climate change will very likely magnify the
2 cumulative impacts of other stresses, such as air and water pollution and habitat destruction
3 caused by human development patterns. For some systems, such as coral reefs, the
4 combined effects of climate change and other stresses are very likely to exceed a critical
5 threshold, bringing large, possibly irreversible impacts.
6 (10) Uncertainties Remain and Surprises Are Expected. Significant uncertainties remain in the
7 science underlying regional climate changes and their impacts. Further research is needed
8 to improve understanding and predictive ability about societal and ecosystem impacts and
9 to provide the public with additional useful information about adaptation strategies.
10 However, it is likely that some aspects and impacts of climate change will be totally
11 unanticipated as complex systems respond to ongoing climate change in unforeseeable
12 ways.
13 For more specific information on the types of effects projected as likely to occur in the United
14 States, the reader is referred to the USGSRC Report (NAST, 2000), several subsidiary regional
15 reports (MARAT, 2000; Yarnal et al., 2000; NERAG, 2001; GLRAG, 2000), and the health
16 assessment report (Bernard, et al., 2001).
17 Findings from the USGCRP (NAST, 2000) report and subsidiary regional reports illustrate
18 well the considerable uncertainties and difficulties in projecting likely climate change impacts on
19 regional or local scales. The findings also reflect well the mixed nature of projected potential
20 climate change impacts (combinations of mostly deleterious, but other possible beneficial
21 effects) for U.S. regions and their variation across the different regions. Difficulties in projecting
22 region-specific climate change impacts are complicated by the need to evaluate potential effects
23 of local- or regional-scale changes in key air pollutants not only on global scale temperature
24 trends but also in terms of potentially more local- or regional- scale impacts on temperature and
25 precipitation patterns. Of much importance for this are varying roles played by atmospheric
26 particles.
27
28 4.5.2.2 Airborne Particle Relationships to Global Warming and Climate Change
29 Atmospheric particles both scatter and absorb incoming solar radiation at visible light
30 wavelengths. The scattering of solar radiation back to space leads to a decrease in transmission
31 of visible radiation to the Earth's surface and, hence, to a decrease in the heating rate of the
April 2002 4-195 DRAFT-DO NOT QUOTE OR CITE
-------
1 surface and the atmosphere. The absorption of either incoming solar radiation or outgoing
2 terrestrial infrared radiation by atmospheric particles results in heating of the lower atmosphere.
3 Interactions of atmospheric particles with electromagnetic radiation from the visible through the
4 infrared spectral regions are responsible for their direct effects on climate, which are the result of
5 the same physical processes responsible for visibility degradation. Visibility reduction is caused
6 by particle scattering in all directions, whereas climate effects result mainly from scattering in the
7 upward direction. The net effect of the above processes can be expressed as a radiative forcing,
8 which is the change in the average net radiation at the top of the troposphere because of a change
9 in solar (shortwave, or visible) or terrestrial (longwave, or infrared) radiation (Houghton et al.,
10 1990). The radiative forcing drives the climate to respond, but because of uncertainties in a
11 number of feedback mechanisms involving climate response, radiative forcing is used as a first-
12 order estimate of the potential importance of various substances. Sulfate particles scatter solar
13 radiation effectively and do not absorb at visible wavelengths, whereas they absorb weakly at
14 infrared wavelengths (IPCC, 2001). Nitrate particles exhibit grossly similar properties. The
15 effects of mineral dust particles are complex; they weakly absorb solar radiation but their overall
16 effect on solar radiation depends on particle size and the reflectivity of the underlying surface.
17 They absorb infrared radiation and thus contribute to greenhouse warming (Tegen et al., 1996).
18 Organic carbon particles mainly reflect solar radiation, whereas elemental carbon and other black
19 carbon particles (e.g., PAHs with H:C ratios of <0.3) are strong absorbers of solar radiation
20 (IPCC, 2001). However, the optical properties of carbonaceous particles are modified if they
21 become coated with water or sulfuric acid. Particles containing black carbon also can exert a
22 direct effect after deposition onto surfaces that are more reflective (e.g., snow and ice). In this
23 case, additional solar radiation is absorbed by the surface; conversely, more reflective particles
24 deposited on a dark surface result in additional solar radiation being reflected back to space.
25 Anthropogenic (Twomey, 1974; Twomey, 1977) and biogenic (Charlson et al., 1987)
26 sulfate particles also exert indirect effects on climate by serving as cloud condensation nuclei,
27 which results in changes in the size distribution of cloud droplets by producing more particles
28 with smaller sizes. The same mass of liquid water in smaller particles leads to an increase in
29 amount of solar radiation that clouds reflect back to space because the total surface area of the
30 cloud droplets is increased. This has been supported by satellite observations indicating that the
31 effective radius of cloud droplets is smaller in the Northern Hemisphere than in the Southern
April 2002 4-196 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hemisphere (Han et al., 1994). Smaller cloud droplets also have a lower probability of
2 precipitating and, thus, have a longer lifetime than larger ones. Although the effects of sulfate
3 have been considered most widely, interactions with other aerosol components also may be
4 important. Novakov and Penner (1993) have provided evidence that carbonaceous particles can
5 modify the nucleation properties of sulfate particles.
6 The amount of solar radiation incident on the earth-atmosphere system, or the solar
7 constant, is 1370 W m"2, or 342.5 W m"2 on a globally averaged basis (calculated by dividing the
8 solar constant by 4). The addition of sulfate and organic carbon as airborne PM results in
9 enhanced scattering and net cooling, whereas the addition of particles containing elemental
10 carbon results in absorption of solar and terrestrial radiation and net heating. The estimated
11 radiative forcing because of the scattering of solar radiation back to space caused mainly by
12 sulfate particles is -0.4 W m"2; -0.2 W m"2 for biomass-burning aerosols; -0.1 W m"2 for fossil
13 fuel organic carbon; and +0.2 W m"2 for fossil fuel black carbon. Uncertainties in these
14 quantities are about a factor of two (IPCC, 200 la). The uncertainty range reflects uncertainties in
15 the emissions of SO2, the amount of SO2 that is oxidized to sulfate, the atmospheric lifetime of
16 sulfate, and the optical properties of the sulfate particles. These values may be compared to the
17 radiative forcing exerted by greenhouse gases of about + 2.4 W m"2, with an uncertainty factor of
18 1.15 from the preindustrial era (ca. 1750) to 2000. Since the beginning of the 20th century, the
19 mean surface temperature of the earth has increased by about 0.6 °C (IPCC, 2001a). Estimates
20 of the indirect effects of particles range from 0 to -2.0 W m"2 (IPCC, 2001a). Because of a lack
21 of quantitative knowledge, no central value could be given. Therefore, on a globally averaged
22 basis, the direct and indirect effects of anthropogenic sulfate particles likely have offset partially
23 the warming effects caused by increases in levels of greenhouse gases (Charlson et al., 1992).
24 Much of the work investigating the effects of particles on climate has focused on sulfate
25 particles. However, particles containing elemental carbon (EC) from fossil fuel combustion and
26 biomass burning or mineral dust may exert radiative forcing, with spatial distributions very
27 different than for sulfate. Tegen et al. (1996) and Tegen and Lacis (1996) used a global scale
28 three-dimensional model to evaluate the radiative forcing caused by mineral dust particles.
29 Tegen and Lacis (1996) found that the sign and the magnitude of the radiative forcing depends on
30 the height distribution of the dust and the effective radius of the particles. In particular, for a dust
31 layer extending from 0 km to 3 km, positive radiative forcing at visible wavelengths is found for
April 2002 4-197 DRAFT-DO NOT QUOTE OR CITE
-------
1 particle radii greater than 1.8 //m, whereas negative forcing is found for smaller particles. They
2 calculated a global mean radiative forcing caused by mineral dust from all sources of 0.14 W m"2
3 and from mineral dust from lands disturbed by human activity of 0.09 W m"2. This value
4 represents a near cancellation between a much larger solar forcing of -0.25 W m"2 and a thermal
5 forcing of 0.34 W m"2. Uncertainty factors could not be estimated for these calculations because
6 they were judged to be largely unknown. Haywood and Shine (1995) estimated a global mean
7 radiative forcing of 0.1 W m"2, with an uncertainty factor >3, caused by the absorption of solar
8 radiation by EC released by fossil fuel combustion. The IPCC (1995) estimated a global mean
9 radiative forcing of -0.1 W m"2 caused by particles produced by biomass burning, with an
10 uncertainty factor of three. The global mean radiative forcing exerted by particles would then be
11 -0.5 W m"2, with an uncertainty of about a factor of 2.4. Figure 4-32 summarizes estimates of
12 global mean radiative forcing exerted by greenhouse gases and various types of particles.
13 Deviations from the global mean values can be very large on the regional scale.
14 For instance, Tegen et al. (1996) found that local radiative forcing exerted by dust raised from
15 disturbed lands ranges from -2.1 W m"2 to 5.5 W m"2 over desert areas and their adjacent seas.
16 The largest regional values of radiative forcing caused by anthropogenic sulfate are about
17 -3 W m"2 in the eastern United States, south central Europe, and eastern China (Kiehl and
18 Briegleb, 1993). These regional maxima in aerosol forcing are at least a factor of 10 greater than
19 their global mean values shown in Figure 4-32. By comparison, regional maxima in forcing by
20 the well-mixed greenhouse gases are only about 50% greater than their global mean value (Kiehl
21 and Briegleb, 1993). Thus, the estimates of local radiative forcing by particles also are large
22 enough to completely cancel the effects of greenhouse gases in many regions and to cause a
23 number of changes in the dynamic structure of the atmosphere that still need to be evaluated.
24 A number of anthropogenic pollutants whose distributions are highly variable are also effective
25 greenhouse absorbers. These gases include O3 and, possibly, HNO3, C2H4, NH3, and SO2, all of
26 which are not commonly considered in radiative forcing calculations (Wang et al. 1976). High
27 ozone values are found downwind of urban areas and areas where there is biomass burning.
28 However, Van Borland et al. (1997) found that there may not be much cancellation between the
29 radiative effects for ozone and for sulfate, because both species have different seasonal cycles
30 and show significant differences in their spatial distribution.
31
April 2002 4-198 DRAFT-DO NOT QUOTE OR CITE
-------
2-
!_ D)
ns .g
2. E
_
0)
Q_
O)
C
o g>
£ |
(D O
.>
"ro
3
n:
1-
-1-
-2H
T
1
Halocarbons
N2O
Aerosols
°n4 c Black ^
_... carbon from
^u2 fossil
Tropospheric fue| Mineral
ozone burning Dust
. T
1
L^J T
T T
1
4-1 ^P
Stratospheric Organic 1
ozone 1 carbon Biomass Aprnt;,
— II. from i • MCI ub1
Sulphate ™™ burning jndire(
Aviation-induced
j^ Solar
Hr1
Land-
3l use
;t (albedo)
fuel effect only
burning
High Medium Medium Low
Very Very Very Very
Low Low Low Low
Very Very Very Very
Low Low Low Low
Level of Scientific Understanding
Figure 4-32. Estimated global mean radiative forcing exerted by gas and various particle
phase species for the year 2000, relative to 1750.
Source: IPCC(2001a).
1 Observational evidence for the climatic effects of particles is sparse. Haywood et al.
2 (1999) found that the inclusion of anthropogenic aerosols results in a significant improvement
3 between calculations of reflected sunlight at the top of the atmosphere and satellite observations
4 in oceanic regions close to sources of anthropogenic PM.
5 Uncertainties in calculating the direct effect of airborne particles arise from a lack of
6 knowledge of their vertical and horizontal variability, their size distribution, chemical
7 composition and the distribution of components within individual particles. For instance,
8 gas-phase sulfur species may be oxidized to form a layer of sulfate around existing particles in
9 continental environments, or they may be incorporated in sea-salt particles (e.g., Li-Jones and
10 Prospero, 1998). In either case, the radiative effects of a given mass of the sulfate will be much
11 lower than if pure sulfate particles were formed. It also must be stressed that the overall radiative
April 2002
4-199
DRAFT-DO NOT QUOTE OR CITE
-------
1 effect of particles at a given location is not simply determined by the sum of effects caused by
2 individual classes of particles because of interactions between particles with different radiative
3 characteristics and with gases.
4 Calculations of the indirect effects of particles on climate are subject to much larger
5 uncertainties than are calculations of their direct effects, reflecting uncertainties in a large
6 number of chemical and microphysical processes in describing the effects of sulfate on the size
7 distribution and number of droplets within a cloud. A complete assessment of the radiative
8 effects of PM will require supercomputer calculations that incorporate the spatial and temporal
9 behavior of particles of varying composition that have been emitted or formed from precursors
10 emitted from different sources. Refining values of model input parameters (such as improving
11 emissions estimates) may be as important as improving the models per se in calculations of direct
12 radiative forcing (Pan et al., 1997) and indirect radiative forcing (Pan et al., 1998) caused by
13 sulfate. However, uncertainties associated with the calculation of radiative effects of particles
14 likely will remain much larger than those associated with well-mixed greenhouse gases.
15 This means that, although on a global scale atmospheric particles likely exert an overall net
16 effect of slowing global warming, much uncertainty would apply to any modeling efforts aimed
17 at projecting net effects on global warming processes, resulting climate change, and any
18 consequent human health or environmental effects because of location-specific increases or
19 decreases in anthropogenic emissions of atmospheric particles or their precursors. For example,
20 any net impacts of regional sulfates in reducing global-climate-change-induced increases in local
21 temperatures may well be offset partially by local surface level heating because of carbonaceous
22 particles from diesel emissions or coal combustion energy generation being deposited on snow or
23 ice covered surfaces or contributing to more rapid evaporation or rainout of water from overhead
24 clouds.
25
26
27 4.6 SUMMARY
28 4.6.1 Particulate Matter Effects on Vegetation and Ecosystems
29 Particulate matter (PM) deposition on vegetation and ecosystems has been defined mainly
30 by size fraction, not by chemical composition, structure, or source. Though size is related to
April 2002 4-200 DRAFT-DO NOT QUOTE OR CITE
-------
1 mode and magnitude of deposition to vegetated landscapes and may be a useful surrogate for
2 chemical composition, the size classes have little specific relevance to vegetation.
3 Deposition of PM on vegetation and ecosystems is not well understood. Atmospheric
4 deposition of particles takes place via both wet and dry processes via three major routes:
5 (1) precipitation and scavenging in which particles are deposited in rain and snow; (2) occult
6 (fog, cloud water, and mist interception); and (3) the much slower dry deposition. All three
7 modes of deposition must be considered when determining inputs to ecosystems or water sheds
8 because each may dominate over specific intervals of time of space.
9 Wet deposition is generally confounded by fewer factors than the other two methods and
10 has been easier to quantify. Total inputs by wet deposition can be significant; however, not all
11 wet deposition involves particle scavenging because gaseous pollutants also dissolve in rain
12 drops during precipitation events. This contribution is obscured because wet deposition is
13 measured simply by chemical analysis of total precipitation collected in clean non-reactive
14 buckets. Wet deposition is largely function of precipitation amount and ambient pollutant
15 concentrations. Surface properties are relatively unimportant except for landform features that
16 alter local distribution of precipitation (orographic effects). Wet deposition is most effective for
17 fine particles of atmospheric (secondary) origin and elements such as cadmium, chromium, lead,
18 nickel, and vanadium.
19 Dry deposition depends more strongly on surface properties, such as micrometeorological
20 roughness, which determine impaction and reentrainment of individual particles, and on particle
21 size distribution in the atmosphere. Vegetation discontinuities, such as forest edges and margins
22 of cultivated fields, may be subject to increased deposition of PM. Dry deposition of
23 atmospheric particles to plant and soil is a much slower processes than either wet or occult
24 deposition and is most effective for coarse particles including primary geologic material, and for
25 elements such as iron and manganese. It is nearly continuous and affects all types of plant parts,
26 including those not currently physiologically active, along with exposed soil and water surfaces,
27 received steady deposits of dry dusts, elemental carbon encrustations, grease films, tarry acidic
28 coatings and heterogeneous secondary particles for from gaseous precursors.
29 Occult deposition is of more restricted occurrence than either of the above. Occult
30 deposition of cloud and fog water droplets containing PM may be determined by both
31 atmospheric and surface features. Formation of fog may accelerate deposition by transforming
April 2002 4-201 DRAFT-DO NOT QUOTE OR CITE
-------
1 fine PM with low deposition velocities, into larger hydrometeors with correspondingly larger
2 deposition velocities.
3 The sources and forms of nitrogen in the atmosphere are poorly studied, and the
4 concentrations are rarely measured, except in precipitation. The influence of aerodynamic
5 diameter is particularly critical for the deposition of nitrogen species because they exist as a wide
6 range of particle sizes in the atmosphere. For example, at many North American sites NO3" is
7 characterized by bimodal distribution, with modes above and below 1 //m. Although the annual
8 deposition for NH4+ is distributed similarly among the fine and coarse particles, particulate NO3"
9 is found predominately in coarse particle fractions. Similar to the pattern for NH4+, the estimated
10 annual deposition of SO4"2 particles occurs in both the fine and coarse particulate fractions. Base
11 cation deposition is virtually restricted to contributions from coarse particles.
12 The ambient concentration of particles, the parameter for which there is most data
13 (Chapter 3), is at best a surrogate indicator of exposure. The amount entering the immediate
14 plant environment, deposition to the plant surfaces or soil in the vicinity of the roots, determines
15 the biological effect.
16 Annual amounts of total heavy metal deposition are highly variable depending on specific
17 forest location and upwind source strength. Depending on climate conditions and topography,
18 fine particles may remain airborne for days to months and may be transported 1,000 to 10,
19 000 km or more from their source. This long-distant transport and subsequent deposition qualify
20 heavy metals as regional- and global-scale air pollutants. Ecosystems immediately downwind of
21 major emissions sources such as power generating, industrial, or urban complexes may receive
22 locally heavy inputs. Mass balance budgets of seven heavy metals (cadmium, copper, iron,
23 manganese, nickel, and zinc) have been determined at the Hubbard Brook Experimental Forest
24 approximately 120 km north of Boston and relatively distant from major source of heavy metals.
25 Investigations of trace metals conducted in roadside, industrial, and urban environments have
26 demonstrated that impressive burdens of particulate heavy metal accumulate on vegetative
27 surfaces. Theory and measurement techniques for wet and occult deposition processes are well
28 advanced. In contrast, dry deposition of particles has remained difficult to measure and to model.
29 Further advances in quantification of PM deposition will require development of improved
30 analytical treatments of dry deposition, and increased chemical speciation of size classed PM.
April 2002 4-202 DRAFT-DO NOT QUOTE OR CITE
-------
1 Human existence on this planet depends on ecosystems and the services and products they
2 provide. Both ecosystem structure and function play an essential role in providing societal
3 benefits. Society derives two types of benefits from the structural aspects of an ecosystem:
4 (1) products with market value such as fish, minerals, forage, forest products, biomass fuels,
5 natural fiber, and many pharmaceuticals, and the genetic resources of valuable species (e.g.,
6 plants for crops and timber and animals for domestication); and (2) the use and appreciation of
7 ecosystem for recreation, aesthetic enjoyment, and study.
8 Ecosystem functions that maintain clean water, pure air, a green earth, and a balance of
9 creatures, are functions that enable humans to survive. They are the dynamics of ecosystems.
10 The benefits they impart include absorption and breakdown of pollutants, cycling of nutrients,
11 binding of soil, degradation of organic waste, maintenance of a balance of gases in the air,
12 regulation of radiation balance, climate, and the fixation of solar energy. Concern has risen in
13 recent years concerning the integrity of ecosystems because there are few ecosystems on Earth
14 today that are not influenced by humans. For this reason, the deposition of PM and its impact on
15 vegetation and ecosystems is of great importance.
16 The PM whose effects on vegetation and ecosystems are considered in this chapter is not a
17 single pollutant but represents a heterogeneous mixture of particles differing in origin, size, and
18 chemical constituents. The effects of exposure to a given mass concentration of PM of particular
19 size (measured as PM10; PM2 5, etc.) may, depending on the particular mix of deposited particles,
20 lead to widely differing phytotoxic responses. This has not been characterized adequately.
21 Atmospheric deposition of particles to ecosystems takes place via both wet and dry
22 processes through the three major routes indicated below.
23 (1) Precipitation scavenging, in which particles are deposited in rain and snow
24 (2) Fog, cloud water, and mist interception
25 (3) Dry deposition, a much slower, yet more continuous removal to surfaces
26 Deposition of heavy metal particles to ecosystems occurs by wet and dry processes. Dry
27 deposition is considered more effective for coarse particles of natural origin and elements such as
28 iron and manganese, whereas wet deposition generally is more effective for fine particles of
29 atmospheric origin and elements such as cadmium, chromium, lead, nickel, and vanadium. The
30 actual importance of wet versus dry deposition, however, is highly variable, depending on the
31 type of ecosystem, location, and elevation.
April 2002 4-203 DRAFT-DO NOT QUOTE OR CITE
-------
1 Deposition of PM on above-ground plant parts can have either a physical and or chemical
2 impact, or both. Particles transferred from the atmosphere to plant surfaces may cause direct
3 effects if they (1) reside on the leaf, twig, or bark surface for an extended period; (2) be taken up
4 through the leaf surface; or (3) are removed from the plant via resuspension to the atmosphere,
5 washing by rainfall, or litter-fall with subsequent transfer to the soil.
6 Chemical effects include excessive alkalinity or acidity. The effects of "inert" PM are
7 mainly physical, whereas the effects of toxic particles are both chemical and physical. The
8 effects of dust deposited on plant surfaces or on soil are more likely to be associated with their
9 chemistry than with the mass of deposited particles and are usually of more importance than any
10 physical effects. The majority of the easily identifiable direct and indirect effects, other than
11 climate-change impacts, occur in severely polluted areas around heavily industrialized point
12 sources such as limestone quarries; cement kilns; and iron; lead, and various smelting factories.
13 Studies of the direct effects of chemical additions to foliage in particulate deposition have found
14 little or no effects of PM on foliar processes; however, both conifers and deciduous species have
15 shown significant effects on leaf surface structures after exposure to simulated acid rain or mist
16 at pH 3.5. Many experimental studies indicate that epicuticular waxes (which function to prevent
17 water loss from plant leaves) can be destroyed by acid rain in a few weeks. This function is
18 particularly crucial in conifers because of the longevity of evergreen foliage.
19 Though there has been no direct evidence of a physiological association between tree injury
20 and exposure to metals, heavy metals have been implicated because their deposition pattern is
21 correlated with forest decline. The role of heavy metals has been indicated by phytochelatin
22 measurements. Phytochelatins are intracellular metal-binding peptides that act as indicator of
23 metal stress. Because they are produced by plants as a response to sublethal concentrations of
24 heavy metals, they can be used to indicate that heavy metals are involved in forest decline.
25 Concentrations of the phytochelatins increased with altitude, as did forest decline, and they also
26 increased across regions showing increased levels of forest injury.
27 Secondary organics formed in the atmosphere have been referred to under the following
28 terms: toxic substances, pesticides, hazardous air pollutants (HAPS), air toxics, semivolatile
29 organic compounds (SOCs), and persistent organic pollutants (POPS). The chemical substances
30 listed under the above headings are not criteria pollutants controlled by NAAQS as cited under
31 CAA Sections 108 and 109 (U.S. Code, 1991), but rather are controlled under CAA Sect. 112,
April 2002 4-204 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hazardous Air Pollutants. Their possible effects in the environment on humans and ecosystems
2 are discussed in many other government documents and publications. They are mentioned in this
3 chapter because, in the atmosphere many of the chemical compounds are partitioned between gas
4 and particle phases and are deposited as particulate matter. As particles, they become airborne
5 and can be distributed over a wide area and impact remote ecosystems. Some of the chemical
6 compounds are of concern to humans because they may reach toxic levels in food chains of both
7 animals and humans, whereas others tend to decrease or maintain the same toxicity as they move
8 through the food chain.
9 An important characteristic of fine particles is their ability to affect the flux of solar
10 radiation passing through the atmosphere directly, by scattering and absorbing solar radiation,
11 and indirectly, by acting as cloud condensation nuclei that, in turn, influence the optical
12 properties of clouds. Regional haze has been estimated to diminish surface solar visible radiation
13 by approximately 8%. Crop yields have been reported as being sensitive to the amount of
14 sunlight received, and crop losses have been attributed to increased airborne particle levels in
15 some areas of the world.
16 The transmission of solar UV-B radiation through the Earth's atmosphere is controlled by
17 ozone, clouds, and particles. The depletion of stratospheric ozone caused by the release of
18 chlorofluorcarbons and other ozone-depleting substances has resulted in heightened concern
19 regarding potentially serious increases in the amount of solar UV-B (SUVB) radiation reaching
20 the Earth's surface. Plant species vary enormously in their response to UV-B exposures, and
21 large differences in response also occur among different genotypes within a species. In general,
22 dicotyledonous plants are more sensitive than monocotyledons from similar environments.
23 In addition, plant responses may differ depending on stage of development. Because plants
24 evolved under the selective pressure of ambient UV-B radiation in sunlight, they have developed
25 adaptive mechanisms. Although inhibition of photosynthesis is a detrimental growth effect,
26 flavonoid synthesis represents acclimation. Plants growing under full light have been shown to
27 be protected against UV-B effects but not when growing under weak visible light. A common
28 adaptation is alteration in leaf transmission properties, which results in attenuation of UV-B in
29 the epidermis before it can reach the leaf interior.
30 Indirect effects of PM are considered of greatest significance because their deposition on
31 the soil has altered nutrient cycling and inhibited nutrient uptake and changed the functioning,
April 2002 4-205 DRAFT-DO NOT QUOTE OR CITE
-------
1 structure and biodiversity of ecosystems. Indirect effects occur through the soil and result from
2 the deposition of heavy metals, nitrates, sulfates, or acidic precipitation and their impact on the
3 soil microbial community. The soil environment is one of the most dynamic sites of biological
4 interaction in nature. Bacteria in the soil are essential components of the nitrogen and sulfur
5 cycles that make these elements available for plant uptake. Fungi form mycorrhizae,
6 a mutualistic symbiotic relationship, that is integral in mediating plant uptake of mineral
7 nutrients. Changes in the soil environment that influence the role of the bacteria and fungi in
8 nutrient cycling and availability determine plant and ecosystem response.
9 Major impacts of PM on soil environments occur through deposition of nitrates and
10 sulfates and the acidifying effect of the FT ion associated with these compounds in wet and dry
11 deposition. Although the soils of most of North American forest ecosystems are nitrogen
12 limited, there are some forests that exhibit severe symptoms of nitrogen saturation. They include
13 the high-elevation, spruce-fir ecosystems in the Appalachian Mountains; the eastern hardwood
14 watersheds at the Fernow Experimental Forest near Parsons, WV; the mixed conifer forest and
15 chaparral watershed with high smog exposure in the Los Angeles Air Basin; the high-elevation
16 alpine watersheds in the Colorado Front Range; and a deciduous forest in Ontario, Canada.
17 Nitrogen saturation results when additions to soil background nitrogen (nitrogen loading)
18 exceed the capacity of plants and soil microorganisms to utilize and retain nitrogen. An
19 ecosystem no longer functions as a sink under these circumstances. Possible ecosystem
20 responses to nitrate saturation, as postulated by Aber and his coworkers, include (1) a permanent
21 increase in foliar nitrogen and reduced foliar phosphorus and lignin because of the lower
22 availability of carbon, phosphorus, and water; (2) reduced productivity in conifer stands caused
23 by disruptions of physiological function; (3) decreased root biomass and increased nitrification
24 and nitrate leaching; (4) reduced soil fertility, the results of increased cation leaching, increased
25 nitrate and aluminum concentrations in streams, and decreased water quality. Saturation implies
26 that some resource other than nitrogen is limiting biotic function. Water and phosphorus for
27 plants and carbon for microorganisms are the resources most likely to be the secondary limiting
28 factors. The appearance of nitrogen in soil solution is an early symptom of excess nitrogen.
29 In the final stage, disruption of forest structure becomes visible.
30 Changes in nitrogen supply can have a considerable impact on an ecosystem's nutrient
31 balance. Increases in soil nitrogen play a selective role. Plant succession patterns and
April 2002 4-206 DRAFT-DO NOT QUOTE OR CITE
-------
1 biodiversity are affected significantly by chronic nitrogen additions in some ecosystems.
2 Long-term nitrogen fertilization studies in both New England and Europe suggest that some
3 forests receiving chronic inputs of nitrogen may decline in productivity and experience greater
4 mortality. For example, long-term fertilization experiments at Mount Ascutney, VT, suggest that
5 declining coniferous forest stands with slow nitrogen cycling may be replaced by deciduous
6 fast-growing forests that cycle nitrogen rapidly. Excess nitrogen inputs to unmanaged heathlands
7 in the Netherlands also have been found to result in nitrophilous grass species replacing slower
8 growing heath species. Over the past several decades, the composition of plants in the forest
9 herb layers had been shifting toward species commonly found on nitrogen-rich areas. It also was
10 observed that the fruiting bodies of mycorrhizal fungi had decreased in number.
11 Notable impacts of excess nitrogen deposition also have been observed with regard to
12 aquatic systems. For example, atmospheric nitrogen deposition into soils in watershed areas
13 feeding into estuarine sound complexes (e.g., the Pamlico Sound of North Carolina) appear to
14 contribute to excess nitrogen flows in runoff (especially during and after heavy rainfall events
15 such as hurricanes). Together with excess nitrogen runoff from agricultural practices or other
16 uses (e.g., fertilization of lawns or gardens), massive influxes of such nitrogen into watersheds
17 and sounds can lead to dramatic decreases in water oxygen and increases in algae blooms that
18 can cause extensive fish kills and damage to commercial fish and sea food harvesting.
19 Acidic deposition has played a major role in soil acidification in some areas of Sweden,
20 elsewhere in Europe, and in eastern North America. Soil acidification and its effects result from
21 deposition of nitrates, sulfates, and associated FT ion. A major concern is that soil acidity will
22 lead to nutrient deficiency. Growth of tree species can be affected when high aluminum-to-
23 nutrient ratios limit uptake of calcium and magnesium and create a nutrient deficiency. Calcium
24 is essential in the formation of wood and the maintenance of cells (the primary plant tissues
25 necessary for tree growth), and it must be dissolved in soil water to be taken up by plants. Acidic
26 deposition can increase aluminum concentrations in soil water by lowering the pH in aluminum-
27 rich soils through dissolution and ion-exchange processes. Aluminum in soil can then be taken
28 up by roots more readily than calcium because of its greater affinity for negatively charged
29 surfaces. Tree species can be adversely affected if altered Ca/Al ratios impair Ca or Mg uptake.
30 Overall, then, PM produced by human activities has the potential to cause the loss of
31 ecosystem biodiversity in ways that reduces the ability of ecosystems to provide the services that
April 2002 4-207 DRAFT-DO NOT QUOTE OR CITE
-------
1 society requires to sustain life. The major impacts of PM on ecosystems are the indirect effects
2 that occur through the soil and affect plant growth, vigor, and reproduction. Mineral nutrient
3 cycling can be altered by the deposition of heavy metals. The deposition of nitrogen and sulfur
4 and the acidifying effects of the two in association with the H+ ion in precipitation also alter
5 biogeochemical cycling, cause soil acidification, alter the Ca/Al ratio, and impact the growth of
6 vegetation and forest trees, in particular. Leaching of nitrates and other minerals through runoff
7 can impact coastal and aquatic wetlands and, thus, influence their ability to produce the products
8 and services necessary for existence of human society.
9
10 4.6.2 Particulate Matter-Related Effects on Visibility
11 Visibility is defined as the degree to which the atmosphere is transparent to visible light
12 and the clarity and color fidelity of the atmosphere. Visual range is the farthest distance a black
13 object can be distinquished against the horizontal sky. Visibility impairment is any humanly
14 perceptible change in visibility. For regulatory purposes, visibility impairment, characterized by
15 light extinction, visual range, contrast, and coloration, is classified into two principal forms:
16 (1) "reasonably attributable" impairment, attributable to a single source or small group of
17 sources, and (2) regional haze, any perceivable change in visibility caused by a combination of
18 many sources over a wide geographical area.
19 Visibility is measured by human observation, light scattering by particles, the light
20 extinction-coefficient and parameters related to the light-extinction coefficient (visual range and
21 deciview scale), the light scattering coefficient, and fine PM concentrations. The air quality
22 within a sight path will affect the illumination of the sight path by scattering or absorbing solar
23 radiation before it reaches the Earth's surface. The rate of energy loss with distance from a beam
24 of light is the light extinction coefficient. The light extinction coefficient is the sum of the
25 coefficients for light absorption by gases (oag), light scattering by gases (osg), light absorption by
26 particles (oap), and light scattering by particles (osp). Atmospheric particles are frequently divided
27 into fine and coarse particles. Corresponding coefficients for light scattering and absorption by
28 fine and coarse particles are osfp and oafp and oscp and oacp, respectively. Visibility within a sight
29 path longer than approximately 100 km (60 mi) is affected by change in the optical properties of
30 the atmosphere over the length of the sight path.
April 2002 4-208 DRAFT-DO NOT QUOTE OR CITE
-------
1 Visibility impairment is associated with airborne particle properties, including size
2 distributions (i.e., fine particles in the 0.1- to 1.0-//m size range) and aerosol chemical
3 composition, and with relative humidity. With increasing relative humidity, the amount of
4 moisture available for absorption by particles increases, thus causing the particles to increase in
5 both size and volume. As the particles increase in size and volume, the light scattering potential
6 of the particles also generally increases. Visibility impairment is greatest in the eastern United
7 States and Southern California. In the eastern United States, visibility impairment is caused
8 primarily by light scattering by sulfate aerosols and, to a lesser extent, by nitrate particles and
9 organic aerosols, carbon soot, and crustal dust. Haziness in the southeastern United States,
10 caused by increased atmospheric sulfate, has increased by ca. 80% since the 1950s and is greatest
11 in the summer months, followed by the spring and fall, and winter. Light scattering by nitrate
12 aerosols is the major cause of visibility impairment in southern California. Nitrates contribute
13 about 40% to the total light extinction in southern California and accounts for 10 to 20% of the
14 total extinction in other U.S. areas.
15 Organic particles are the second largest contributors to light extinction in most U.S. areas.
16 Organic carbon is the greatest cause of light extinction in the Pacific Northwest, Oregon, Idaho,
17 and Montana, accounting for 40 to 45% of the total extinction. Also, organic carbon contributes
18 between 15 to 20% to the total extinction in most of the western United States and 20 to 30% in
19 the remaining U.S. areas.
20 Coarse mass and soil, primarily considered "natural extinction", is responsible for some of
21 the visibility impairment in northern California and Nevada, Oregon, southern Idaho, and
22 western Wyoming. Dust transported from southern California and the subtropics has been
23 associated with regional haze in the Grand Canyon and other southwestern U.S. class I areas.
24
25 4.6.3 Particulate Matter-Related Effects on Materials
26 Building materials (metals, stones, cements, and paints) undergo natural weathering
27 processes from exposure to environmental elements (wind, moisture, temperature fluctuations,
28 sun light, etc.). Metals form a protective film that protects against environmentally induced
29 corrosion. The natural process of metal corrosion from exposure to natural environmental
30 elements is enhanced by exposure to anthropogenic pollutants, in particular SO2, rendering the
31 protective film less effective.
April 2002 4-209 DRAFT-DO NOT QUOTE OR CITE
-------
1 Dry deposition of SO2 enhances the effects of environmental elements on calcereous stones
2 (limestone, marble, and cement) by converting calcium carbonate (calcite) to calcium sulfate
3 dihydrate (gypsum). The rate of deterioration is determined by the SO2 concentration, the stone's
4 permeability and moisture content, and the deposition rate; however, the extent of the damage to
5 stones produced by the pollutant species apart from the natural weathering processes is uncertain.
6 Sulfur dioxide also has been found to limit the life expectancy of paints by causing discoloration
7 and loss of gloss and thickness of the paint film layer.
8 A significant detrimental effect of particle pollution is the soiling of painted surfaces and
9 other building materials. Soiling changes the reflectance of a material from opaque and reduces
10 the transmission of light through transparent materials. Soiling is a degradation process that
11 requires remediation by cleaning or washing, and, depending on the soiled surface, repainting.
12 Available data on pollution exposure indicates that particles can result in increased cleaning
13 frequency of the exposed surface and may reduce the life usefulness of the material soiled.
14 Attempts have been made to quantify the pollutants exposure levels at which materials damage
15 and soiling have been perceived. However, to date, insufficient data are available to advance our
16 knowledge regarding perception thresholds with respect to pollutant concentration, particle size,
17 and chemical composition.
18
19 4.6.4 Effects of Atmospheric Particulate Matter on the Transmission of
20 Solar Ultraviolet Radiation and Global Warming Processes
21 Extensive potential future impacts on human health and the environment are projected to
22 occur because of increased transmission of solar ultraviolet radiation (UV-B) through the Earth's
23 atmosphere, secondary to stratospheric ozone depletion resulting from anthropogenic emissions
24 of chlorofluorcarbons (CFCs), halons, and certain other gases. However, the estimation of the
25 likely future extent of detrimental effects caused by increased penetration of solar UV-B to the
26 Earth's surface is complicated by atmospheric particle effects, which vary depending on size and
27 composition of particles that can differ substantially over different geographic areas and from
28 season to season over the same area. Also, atmospheric particles greatly complicate projections
29 of future trends in global warming processes because of emissions of greenhouse gases;
30 consequent increases in global mean temperature, and resulting changes in regional and local
April 2002 4-210 DRAFT-DO NOT QUOTE OR CITE
-------
1 weather patterns; and mainly deleterious (but some beneficial) location-specific human health
2 and environmental impacts.
3 The physical processes (i.e., scattering and absorption) responsible for airborne particle
4 effects on transmission of solar ultraviolet and visible radiation are the same as those responsible
5 for visibility degradation. Scattering of solar radiation back to space and absorption of solar
6 radiation determine the effects of an aerosol layer on solar radiation. The transmission of solar
7 UV-B radiation is affected strongly by atmospheric particles. Measured attenuations of UV-B
8 under hazy conditions range up to 37% of the incoming solar radiation. Measurements relating
9 variations in PM mass directly to UV-B transmission are lacking. Particles also can affect the
10 rates of photochemical reactions occurring in the atmosphere. Depending on the amount of
11 absorbing substances in the particles, photolysis rates either can be increased or decreased.
12 In addition to direct climate effects through the scattering and absorption of solar radiation,
13 particles also exert indirect effects on climate by serving as cloud condensation nuclei, thus
14 affecting the abundance and vertical distribution of clouds. The direct and indirect effects of
15 particles appear to have significantly offset the global warming effects caused by the buildup of
16 greenhouse gases because the onset of the Industrial Revolution, on a globally averaged basis.
17 However, because the lifetime of particles is much shorter than that required for complete mixing
18 within the Northern Hemisphere, the climate effects of particles generally are felt much less
19 homogeneously than are the effects of long-lived greenhouse gases.
20 Any effort to model the impacts of local alterations in particle concentrations on projected
21 global climate change or consequent local and regional weather patterns would be subject to
22 considerable uncertainty. This also would be the case for any projections of impacts of location-
23 specific airborne PM alterations on potential human health or environmental effects associated
24 with either increased atmospheric transmission of solar UV radiation or global warming
25 secondary to accumulation of stratospheric ozone-depleting substances or "greenhouse gases."
April 2002 4-211 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2
3 Abdullah, U. M; Iqbal, M. Z. (1991) Response of automobile, stone and cement paniculate matters on stomatal
4 clogging of plants. Geobios (Jodhpur, India) 18: 196-202.
5 Aber, J. D.; Nadelhoffer, K. J.; Steudler, P.; Melillo, J. M. (1989) Nitrogen saturation in northern forest ecosystems:
6 excess nitrogen from fossil fuel combustion may stress the biosphere. Bioscience 39: 378-386.
7 Aber, J.; McDowell, W.; Nadelhoffer, K.; Magill, A.; Berntson, G.; Kamakea, M.; McNulty, S.; Currie, W.;
8 Rustad, L.; Fernandez, I. (1998) Nitrogen saturation in temperate forest ecosystems. BioScience 48: 921-934.
9 Ainsleigh, H. G. (1993) Beneficial effects of sun exposure on cancer mortality. Prev. Med. 22: 132-140.
10 Alexander, M. (1977) Introduction to soil microbiology. 2nd ed. New York, NY: John Wiley & Sons; pp. 225-230;
11 239-246; 272-286; 355-356.
12 Allen, M. F. (1991) The ecology of mycorrhizae. Cambridge, MA: Cambridge University Press.
13 Allen, A. G.; Harrison, R. M.; Nicholson, K. W. (1991) Dry deposition of fine aerosol to a short grass surface.
14 Atmos. Environ. Part A 25: 2671-2676.
15 Allen, E. R.; Cabrera, N.; Kim, J.-C. (1994) Atmospheric deposition studies in rural forested areas of the
16 southeastern United States. Presented at: 87th annual meeting & exhibition of the Air & Waste Management
17 Association; June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no.
18 94-WP88.02.
19 Allen, E. B.; Padgett, P. E.; Bytnerowicz, A.; Minich, R. (1998) Nitrogen deposition effects on coastal sage
20 vegetation of southern California. USDA Forest Service Gen. Tech. Rep. PSW-GTR-166, pp. 131-139.
21 Altshuller, A. P.; Linthurst, R. A., eds. (1984) Acidic deposition phenomenon and its effects: critical assessment
22 review papers; volume I, atmospheric sciences; volume II, effects sciences. Washington, DC: U.S.
23 Environmental Protection Agency, Office of Research and Development; EPA report nos.
24 EPA/600/8-83/016aF-bF. Available from: NTIS, Springfield, VA; PB85-100030 and PB85-100048.
25 Andersen, C. P.; Rygiewicz, P. T. (1991) Stress interactions and mycorrhizal plant response: understanding carbon
26 allocation priorities. Environ. Pollut. 73: 217-244.
27 Anderson, J. B.; Baumgardner, R. E.; Mohnen, V. A.; Bowser, J. J. (1999) Cloud chemistry in the eastern United
28 States, as sampled from three high-elevation sites along the Appalachian Mountains. Atmos. Environ.
29 33:5105-5114.
30 Aneja, V. P.; Murthy, A. B. (1994) Monitoring deposition of nitrogen-containing compounds in a high-elevation
31 forest canopy. Air Waste 44: 1109-1115.
32 Antonovics, J.; Bradshaw, A. D.; Turner, R. G. (1971) Heavy metal tolerance in plants. In: Cragg, J. B., ed.
33 Advances in ecological research. Volume 7. London, United Kingdom: Academic Press; pp. 1-85.
34 April, R.; Newton, R. (1992) Mineralogy and mineral weathering. In: Johnson, D. W.; Lindberg, S. E., eds.
3 5 Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
36 Springer-Verlag; pp. 378-425. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
37 Ecological studies: analysis and synthesis: v. 91).
38 Arnott, W. P.; Moosmtiller, H.; Rogers, C. F.; Jin, T.; Bruch, R. (1999) Photoacoustic spectrometer for measuring
39 light absorption by aerosol; instrument description. Atmos. Environ. 33: 2845-2852.
40 Askey, A.; Lyon, S. B.; Thompson, G. E.; Johnson, J. B.; Wood, G. C.; Sage, P. W.; Cooke, M. J. (1993) The effect
41 of fly-ash particulates on the atmospheric corrosion of zinc and mild steel. Corros. Sci. 34: 1055-1081.
42 Ausset, P.; Bannery, F.; Del Monte, M.; Lefevre, R. A. (1998) Recording of pre-industrial atmospheric environment
43 by ancient crusts on stone monuments. Atmos. Environ. 32: 2859-2863.
44 Ayensu, E.; Van R. Claasen, D.; Collins, M.; Dearing, A.; Fresco, L.; Gadgil, M.; Gitay, H.; Glaser, G.; Juma, C.;
45 Krebs, J.; lenton, R.; Lubchenco, J.; McNeely, J. A.; Mooney, H. A.; Pinstrup-Andersen, P.; Ramos, M.;
46 Raven, P.; Reid, W. V; Samper, C.; Sarukhan, J.; Schei, P.; Tundisi, J. G.; Watson, R. T.; Guanhua, X.;
47 Zakri, A. H. (1999) International ecosystem assessment. Science 286: 685-686.
48 Babich, H.; Stotzky, G. (1978) Effects of cadmium on the biota: influence of environmental factors. In: Permian, D.,
49 ed. Advances in applied microbiology, volume 23. New York, NY: Academic Press; pp. 55-117.
50 Bache, D. H. (1979a) Particle transport within plant canopies -1. a framework for analysis. Atmos. Environ.
51 13:1257-1262.
52 Bache, D. H. (1979b) Paniculate transport within plant canopies - II. prediction of deposition velocities.
53 Atmos. Environ. 13: 1681-1687.
April 2002 4-212 DRAFT-DO NOT QUOTE OR CITE
-------
1 Baedecker, P. A.; Edney, E. O.; Moran, P. J.; Simpson, T. C.; Williams, R. S. (1991) Effects of acidic deposition on
2 materials. In: Irving, P. M, ed. Acidic deposition: state of science and technology, volume III: terrestrial,
3 materials, health and visibility effects. Washington, DC: The U.S. National Acid Precipitation Assessment
4 Program. (State of science and technology report no. 19).
5 Baedecker, P. A.; Reddy, M. M.; Reimann, K. J.; Sciammarella, C. A. (1992) Effects of acidic deposition on the
6 erosion of carbonate stone-experimental results from the U.S. National Acid Precipitation Assessment
7 Program (NAPAP). Atmos. Environ. Part B 26: 147-158.
8 Baldocchi, D. (1988) A multi-layer model for estimating sulfur dioxide deposition to a deciduous oak forest canopy.
9 Atmos. Environ. 22: 869-884.
10 Baron, J. S.; Rueth, H. M.; Wolfe, A. M.; Nydick, K. R.; Allstott, E. J.; Minear, J. T.; Moraska, B. (2000)
11 Ecosystem responses to nitrogen deposition in the Colorado front range. Ecosystems 3: 352-368.
12 Barone, J. B.; Cahill, T. A.; Eldred, R. A.; Flocchini, R. G.; Shadoan, D. J.; Dietz, T. M. (1978) A multivariate
13 statistical analysis of visibility degradation at four California cities. Atmos. Environ. 12: 2213-2221.
14 Barrie, L. A.; Anlauf, K.; Wiebe, H. A.; Fellin, P. (1984) Acidic pollutants in air and precipitation at selected rural
15 locations in Canada. In: Hicks, B. B., ed. Deposition both wet and dry: [papers presented at the American
16 Chemical Society acid rain symposium]; March 1982; Las Vegas, NV. Boston, MA: Butterworth Publishers;
17 pp. 15-35. (Teasley, J. L, ed. Acid precipitation series: v. 4).
18 Barton, K. (1958) The influence of dust on atmospheric corrosion of metals. Werkst. Korros. 8/9: 547-549.
19 Beier, C. (1991) Separation of gaseous and paniculate dry deposition of sulfur at a forest edge in Denmark.
20 J. Environ. Qual. 20: 460-466.
21 Beier, C.; Gundersen, P.; Rasmussen, L. (1992) A new method for estimation of dry deposition of particles based
22 on throughfall measurements in a forest edge. Atmos. Environ. Part A 26: 1553-1559.
23 Bellan, L. M.; Salmon, L. G.; Cass, G. R. (2000) A study on the human ability to detect soot deposition onto works
24 of art. Environ. Sci. Technol. 34: 1946-1952.
25 Bennett, J. H.; Hill, A. C. (1973) Absorption of gaseous air pollutants by a standardized plant canopy. J. Air Pollut.
26 Control Assoc. 23: 203-206.
27 Bennett, J. H.; Hill, A. C. (1975) Interactions of air pollutants with canopies of vegetation. In: Mudd, J. B.;
28 Kozlowski, T. T., eds. Responses of plants to air pollution. New York, NY: Academic Press; pp. 273-306.
29 Bernard, S. M.; McGeehin, M. G.; Patz, J. A., eds. (2001) Health: the potential consequences of climate variability
30 and change. Environ. Health Perspect. 109(suppl. 2): 175-233.
31 Bewley, R. J. F.; Parkinson, D. (1984) Effects of sulphur dioxide pollution on forest soil microorganisms. Can. J.
32 Microbiol. 30: 179-185.
33 Biefer, G. J. (1981) Atmospheric corrosion of steel in the Canadian Arctic. Mater. Perform. 20: 16-19.
34 Billings, W. D. (1978) Plants and the ecosystem. 3rd ed. Belmont, CA: Wadsworth Publishing Company, Inc.;
35 pp. 1-62, 83-108.
36 Binenko, V. L; Harshvardhan, H. (1993) Aerosol effects in radiation transfer. In: Jennings, S. G., ed.
37 Aerosol effects on climate. Tucson, AZ: University of Arizona Press; pp. 190-232.
38 Binkley, D.; Valentine, D.; Wells, C.; Valentine, U. (1989) An empirical analysis of the factors contributing to
39 20-year decrease in soil pH in an old-field plantation of loblolly pine. Biogeochemistry 8: 39-54.
40 Bjorn, L. O. (1996) Effects of ozone depletion and increased UV-B on terrestrial ecosystems. Int. J. Environ. Stud.
41 51:217-243.
42 Blackburn, T. E.; Bairai, S. T.; Stedman, D. H. (1992) Solar photolysis of ozone to singlet D oxygen atoms.
43 J. Geophys. Res. 97: 10,109-10,117.
44 Blaschke, H. (1990) Mycorrhizal populations and fine root development on Norway spruce exposed to controlled
45 doses of gaseous pollutants and simulated acidic rain treatments. Environ. Pollut. 68: 409-418.
46 Bobbink, R. (1998) Impacts of tropospheric ozone and airborne nitrogenous pollutants on natural and semi-natural
47 ecosystems: a commentary. New Phytol. 139: 161-168.
48 Bobbink, R.; Hornung, M.; Roelofs, J. G. M. (1998) The effects of air-borne nitrogen pollutants on species diversity
49 in natural and semi-natural European vegetation. J. Ecol. 86: 717-738.
50 Bolle, H.-J.; Seiler, W.; Bolin, B. (1986) Other greenhouse gases and aerosols: assessing their role for atmospheric
51 radiative transfer. In: Bolin, B.; Doos, B. R.; Jager, J.; Warrick, R. A., eds. The greenhouse effect, climatic
52 change, and ecosystems. Chichester, United Kingdom: John Wiley and Sons; pp. 157-203. (SCOPE 29).
53
April 2002 4-213 DRAFT-DO NOT QUOTE OR CITE
-------
1 Bondietti, E. A.; McLaughlin, S. B. (1992) Evidence of historical influences of acidic deposition on wood and soil
2 chemistry. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient cycling: a
3 synthesis of the integrated forest study. New York, NY: Springer-Verlag: pp. 358-377. (Billings, W. D.;
4 Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological studies analysis and synthesis: v. 91).
5 Bowden, R. D.; Geballe, G. T.; Bowden, W. B. (1989) Foliar uptake of 15N from simulated cloud water by red
6 spruce (Picea rubens) seedlings. Can. J. For. Res. 19: 382-386.
7 Bowman, W. D. (2000) Biotic controls over ecosystem response to environmental change in alpine tundra of the
8 Rocky Mountains. Ambio 29: 396-400.
9 Bowman, W. D.; Steltzer, H. (1998) Positive feedbacks to anthropogenic nitrogen deposition in Rocky Mountain
10 alpine tundra. Ambio 27: 514-517.
11 Boyce, S. G. (1954) The salt spray community. Ecol. Monogr. 24: 29-67.
12 Bradley, R.; Burt, A. J.; Read, D. J. (1981) Mycorrhizal infection and resistance to heavy metal toxicity in Calluna
13 vulgaris. Nature (London) 292: 335-337.
14 Brandt, C. J.; Rhoades, R. W. (1972) Effects of limestone dust accumulation on composition of a forest community.
15 Environ. Pollut. 3:217-225.
16 Brandt, C. J.; Rhoades, R. W. (1973) Effects of limestone dust accumulation on lateral growth of forest trees.
17 Environ. Pollut. 4: 207-213.
18 Brook, J. R.; Zhang, L.; Di-Giovanni, F.; Padro, J. (1999) Description and evaluation of a model of deposition
19 velocities for routine estimates of air pollutant dry deposition over North America. Part I: model
20 development. Atmos. Environ. 33: 5037-5051.
21 Briihl, C.; Crutzen, P. J. (1988) On the disproportionate role of tropospheric ozone as a filter against solar UV-B
22 radiation. Geophys. Res. Lett. 16: 703-706.
23 Brumme, R.; Leimcke, U.; Matzner, E. (1992) Interception and uptake of NH4 and NO3 from wet deposition by
24 above-ground parts of young beech (Fagus silvatica L.) trees. Plant Soil 142: 273-279.
25 Brunei, J.; Diekmann, M.; Falkengren-Grerup, U. (1998) Effects of nitrogen deposition on field layer vegetation in
26 south Swedish oak forests. Environ. Pollut. (Oxford, U.K.) 102(suppl. 1): 35-40.
27 Bugini, R.; Tabasso, M. L.; Realini, M. (2000) Rate of formation of black crusts on marble. A case study. J. Cult.
28 Heritage 1: 111-116.
29 Businger, J. A. (1986) Evaluation of the accuracy with which dry deposition can be measured with current
30 micrometeorological techniques. J. Clim. Appl. Meteorol. 25: 1100-1124.
31 Butlin, R. N.; Coote, A. T.; Devenish, M.; Hughes, I. S. C.; Hutchens, C. M.; Irwin, J. G.; Lloyd, G. O.; Massey,
32 S. W.; Webb, A. H.; Yates, T. J. S. (1992a) Preliminary results from the analysis of metal samples from the
33 National Materials Exposure Programme (NMEP). Atmos. Environ. Part B 26: 199-206.
34 Butlin, R. N.; Coote, A. T.; Devenish, M.; Hughes, I. S. C.; Hutchens, C. M.; Irwin, J. G.; Lloyd, G. O.; Massey,
35 S. W.; Webb, A. H.; Yates, T. J. S. (1992b) Preliminary results from the analysis of stone tablets from the
36 National Materials Exposure Programme (NMEP). Atmos. Environ. Part B 26: 189-198.
37 Bytnerowicz, A.; Fenn, M. E. (1996) Nitrogen deposition in California forests: a review. Environ. Pollut.
38 92: 127-146.
39 Bytnerowicz, A.; Miller, P. R.; Olszyk, D. M. (1987a) Dry deposition of nitrate, ammonium and sulfate to a
40 Ceanothus crassifolius canopy and surrogate surfaces. Atmos. Environ. 21: 1749-1757.
41 Bytnerowicz, A.; Miller, P. R.; Olszyk, D. M.; Dawson, P. J.; Fox, C. A. (1987b) Gaseous and paniculate air
42 pollution in the San Gabriel Mountains of southern California. Atmos. Environ. 21: 1805-1814.
43 Cahill, T. A.; Eldred, R. A.; Wakabayashi, P. H. (1996) Trends in fine particle concentrations at Great Smoky
44 Mountains National Park. For presentation at: Annual meeting of the Air & Waste Management Association;
45 June; Nashville, TN. Pittsburgh, PA: Air & Waste Management Association; submitted paper no.
46 96-MP1A.05.
47 Campbell, G. G.; Schurr, G. G.; Slawikowski, D. E.; Spence, J. W. (1974) Assessing air pollution damage to
48 coatings. J. Paint Technol. 46: 59-71.
49 Camuffo, D. (1995) Physical weathering of stones. In: Saiz-Jimenez, C., ed. The deterioration of monuments:
50 proceedings of the 2nd international symposium on biodeterioration and biodegradation; January 1994;
51 Sevilla, Spain. Sci. Total Environ. 167: 1-14.
52 Cape, J. N. (1993) Direct damage to vegetation caused by acid rain and polluted cloud: definition of critical levels
53 for forest trees. Environ. Pollut. 82: 167-180.
54 Cape, J. N.; Sheppard, L. J.; Fowler, D.; Harrison, A. F.; Parkinson, J. A.; Dao, P.; Paterson, I. S. (1992)
55 Contribution of canopy leaching to sulphate deposition in a Scots pine forest. Environ. Pollut. 75: 229-236.
April 2002 4-214 DRAFT-DO NOT QUOTE OR CITE
-------
1 Carey, W. F. (1959) Atmospheric deposits in Britain: a study of dinginess. Int. J. AirPollut. 2: 1-26.
2 Ceccotti, S. P.; Messick, D. L. (1997) A global review of crop requirements, supply, and environmental impact on
3 nutrient sulphur balance. In: Cram, W. J.; De Kok, L. J.; Stulen, L; Brunold, C.; Rennenberg, H., eds.
4 Sulphur metabolism in higher plants. Leiden, The Netherlands: Backhuys Publishers; pp. 155-163.
5 Chamberlain, A. C. (1975) The movement of particles in plant communities. In: Monteith, J. L., ed. Vegetation and
6 the atmosphere: v. 1, principles. London, United Kingdom: Academic Press; pp. 155-203.
7 Chamberlain, A. C.; Little, P. (1981) Transport and capture of particles by vegetation. In: Grace, J.; Ford, E. D.;
8 Jarvis, P. G., eds. Plants and their atmospheric environment: the 21st symposium of the British Ecological
9 Society; 1979; Edinburgh, United Kingdom. Oxford, United Kingdom: Blackwell Scientific Publications;
10 pp. 147-173.
11 Chameides, W. L.; Yu, H.; Liu, S. C.; Bergin, M; Zhou, X.; Mearns, L.; Wang, G.; Kiang, C. S.; Saylor, R. D.;
12 Luo, C.; Huang, Y.; Steiner, A.; Giorgi, F. (1999) Case study of the effects of atmospheric aerosols and
13 regional haze on agriculture: an opportunity to enhance crop yields in China through emission controls?
14 Proc. Natl. Acad. Sci. U. S. A. 96: 13626-13633.
15 Chapin, F. S., III. (1991) Integrated responses of plants to stress: a centralized system of physiological responses.
16 Bioscience 41: 29-36.
17 Chapin, F. S., Ill; Ruess, R. W. (2001) The roots of the matter. Nature (London) 411: 749-752.
18 Chapin, F. S., Ill; Bloom, A. J.; Field, C. B.; Waring, R. H. (1987) Plant responses to multiple environmental
19 factors. BioScience 37: 49-57.
20 Chapin, F. S., Ill; Walker, B. H.; Hobbs, R. J.; Hooper, D. U.; Lawton, J. H.; Sala, O. E.; Tilman, D. (1997)
21 Biotic control over functioning ecosystems. Science 277: 500-504.
22 Chapin, F. S., Ill; Sala, O. E.; Burke, I. C.; Grime, J. P.; Hooper, D. U.; Lauenroth, W. K.; Lombard, A.; Mooney,
23 H. A.; Mosier, A. R.; Naeem, S.; Pacala, S. W.; Roy, J.; Steffen, W. L.; Tilman, D. (1998) Ecosystem
24 consequences of changing biodiversity. BioScience 48: 45-52.
25 Charlson, R. J.; Lovelock, J. E.; Andreae, M. O.; Warren, S. G. (1987) Oceanic phytoplankton, atmospheric
26 sulfphur, cloud albedo and climate. Nature (London) 326: 655-661.
27 Charlson, R. J.; Schwartz, S. E.; Hales, J. M.; Cess, R. D.; Coakley, J. A., Jr.; Hansen, J. E.; Hofmann, D. J. (1992)
28 Climate forcing by anthropogenic aerosols. Science (Washington, DC) 255: 423-430.
29 Chevone, B. I.; Herzfeld, D. E.; Krupa, S. V.; Chappelka, A. H. (1986) Direct effects of atmospheric sulfate
30 deposition on vegetation. J. Air Pollut. Control Assoc. 36: 813-815.
31 Choudhury, B. J. (1987) Relationships between vegetation indices, radiation absorption, and net photosynthesis
32 evaluated by a sensitivity analysis. Remote Sens. Environ. 22: 209-233.
33 Chughtai, A. R.; Williams, G. R.; Atteya, M. M. O.; Miller, N. J.; Smith, D. M. (1999) Carbonaceous particle
34 hydration. Atmos. Environ. 33: 2679-2687.
35 Clarke, J. F.; Edgerton, E. S.; Martin, B. E. (1997) Dry deposition calculations for the clean air status and trends
36 network. Atmos. Environ. 31: 3667-3678.
37 Clarkson, D. T.; Sanderson, J. (1971) Inhibition of the uptake and long-distance transport of calcium by aluminum
38 and other polyvalent cations. J. Exp. Bot. 22: 837-851.
39 Clevenger, T. E.; Saiwan, C.; Koirtyohann, S. R. (1991) Lead speciation of particles on air filters collected in the
40 vicinity of a lead smelter. Environ. Sci. Technol. 25: 1128-1133.
41 Clough, W. S. (1975) The deposition of particles on moss and grass surfaces. Atmos. Environ. 9: 1113-1119.
42 Cocker, D. R., Ill; Whitlock, N. E.; Flagen, R. C.; Seinfield, J. H. (2001) Hygroscopic properties of Pasadena,
43 California aerosol. Aerosol Sci. Technol. 35: 637-647.
44 Cole, D. W.; Johnson, D. W. (1977) Atmospheric sulfate additions and cation leaching in a Douglas fir ecosystem.
45 Water Resour. Res. 13: 313-317.
46 Collett, J. L., Jr.; Hoag, K. J.; Sherman, D. E.; Bator, A.; Richards, L. W. (1999) Spatial and temporal variations in
47 San Joaquin Valley fog chemistry. Atmos. Environ. 33: 129-140.
48 Cooke, R. U.; Gibbs, G. B. (1994) Crumbling heritage? Studies of stone weathering in polluted atmospheres.
49 Atmos. Environ. 28: 1355-1356.
50 Costanza, R.; d'Arge, R.; De Groot, R.; Farber, S.; Grasso, M.; Harmon, B.; Limburg, K.; Naeem, S.; O'Neill, R. V.;
51 Paruelo, J.; Raskin, R. G.; Sutton, P.; Van Den Belt, M. (1997) The value of the world's ecosystem services
52 and natural capital. Nature (London) 387: 253-259.
53 Cotrufo, M. F.; De Santo, A. V.; Alfani, A.; Bartoli, G.; De Cristofaro, A. (1995) Effects of urban heavy metal
54 pollution on organic matter decomposition in Quercus ilexL. woods. Environ. Pollut. 89: 81-87.
April 2002 4-215 DRAFT-DO NOT QUOTE OR CITE
-------
1 Cowling, J. E.; Roberts, M. E. (1954) Paints, varnishes, enamels, and lacquers. In: Greathouse, G. A.; Wessel, C. J.,
2 eds. Deterioration of materials: causes and preventive techniques. New York, NY: Reinhold Publishing
3 Corporation; pp. 596-645.
4 Cramer, S. D.; McDonald, L. G.; Bhagia, G.; Flinn, D. R.; Linstrom, P. J.; Carter, J. P. (1989) Effects of acid
5 deposition on the atmospheric corrosion of structural metals. Research Triangle Park, NC: U.S.
6 Environmental Protection Agency; internal report. Incorporated in: Irving, P. M., ed. Acidic Deposition:
7 State of Science and Technology. Volume III: Terrestrial, Materials, Health and Visibility Effects; report 19,
8 section 2. Washington, DC: National Acid Precipitation Assessment Program.
9 Creighton, P. J.; Lioy, P. J.; Haynie, F. H.; Lemmons, T. J.; Miller, J. L.; Gerhart, J. (1990) Soiling by atmospheric
10 aerosols in an urban industrial area. J. Air Waste Manage. Assoc. 40: 1285-1289.
11 Cronan, C. S.; Grigal, D. F. (1995) Use of calcium/aluminum ratios as indicators of stress in forest ecosystems.
12 J. Environ. Qual. 24: 209-226.
13 Cullen, J. J.; Neale, P. J. (1997) Biological weighting functions for describing the effects of ultraviolet radiation on
14 aquatic ecosystems. In: Hader, D.-P., ed. The effects of ozone depletion on aquatic ecosystems. Austin, TX:
15 Academic Press, pp. 97-118.
16 Dahlgren, R. A. (1994) Soil acidification and nitrogen saturation from weathering of ammonium-bearing rock.
17 Nature 368: 838-841.
18 Daily, G. C. (1997) Introduction: what are ecosystem services? In: Daily, G. C., ed. Nature's services: societal
19 dependence on natural ecosystems. Washington, DC: Island Press; pp. 1-10.
20 Daily, G. C.; Ehrlich, P. R. (1999) Managing Earth's ecosystems: an interdisciplinary challenge. Ecosystems
21 2:277-280.
22 Darley, E. F. (1966) Studies on the effect of cement-kiln dust on vegetation. J. Air Pollut. Control Assoc.
23 16: 145-150.
24 Dasch, J. M. (1987) Measurement of dry deposition to surfaces in deciduous and pine canopies. Environ. Pollut.
25 44:261-277.
26 Dasch, J. M.; Cadle, S. H. (1985) Wet and dry deposition monitoring in southeastern Michigan. Atmos. Environ.
27 19:789-796.
28 Dasch, J. M.; Cadle, S. H. (1986) Dry deposition to snow in an urban area. Water Air Soil Pollut. 29: 297-308.
29 Dassler, H.-G.; Ranft, H.; Rehn, K.-H. (1972) Zur Widerstandsfahigkeit von Geholzen gegentiber
30 Fluorverbindungen und Schwefeldioxid [The susceptibility of woody plants exposed to fluorine compounds
31 and SO2]. Flora (Jena) 161: 289-302.
32 Davidson, C. L; Friedlander, S. K. (1978) A filtration model for aerosol dry deposition: application to trace metal
33 deposition from the atmosphere. J. Geophys. Res. C: Oceans Atmos. 83: 2343-2352.
34 Davidson, C. L; Wu, Y.-L. (1989) Dry deposition of trace elements. In: Pacyna, J.M.; Ottar, B., eds. Control and
3 5 fate of atmospheric trace metals: proceedings of the NATO advanced research workshop on fate and control
36 of toxic metals in the atmosphere; September 1988; Oslo, Norway. Boston, MA: Kluwer Academic
37 Publishers; pp. 147-202. (NATO ASI series C, mathematical and physical sciences: v. 268).
38 Davidson, C. L; Wu, Y.-L. (1990) Dry deposition of particles and vapors. In: Lindberg, S. E.; Page, A. L.; Norton,
39 S. A., eds. Acidic precipitation: v. 3, sources, deposition, and canopy interactions. New York, NY:
40 Springer-Verlag; pp. 103-216.
41 Davidson, C. L; Miller, J. M.; Pleskow, M. A. (1982) The influence of surface structure on predicted particle dry
42 deposition to natural grass canopies. Water Air Soil Pollut. 18: 25-43.
43 Davidson, C. L; Tang, W.; Finger, S.; Etyemezian, V.; Striegel, M. F.; Sherwood, S. I. (2000) Soiling patterns on a
44 tall limestone building: changes over 60 years. Environ. Sci. Technol. 34: 560-565.
45 Davis, G. D.; Shaw, B. A.; Arah, C. O.; Fritz, T. L.; Moshier, W. C.; Simpson, T. C.; Moran, P. J.; Zankel, K. L.
46 (1990) Effects of SO2 deposition on painted steel surfaces. Surf. Interface Anal. 15: 107-112.
47 Day, D. E.; Malm, W. C.; Kreidenweis, S. M.; Tree, R. M. (1997) Aerosol light scattering measurements: a
48 comparison of differently configured optec nephelometers. In: Visual air quality: aerosols and global
49 radiation balance: proceedings of a specialty conference sponsored by the Air & Waste Management
50 Association and the American Geophysical Union; September; Bartlett, NH. Pittsburgh, PA: Air & Waste
51 Management Association; pp. 952-962. (A&WMA publication VIP-76).
52 DeHayes, D. H.; Schaberg, P. G.; Hawley, G. J.; Strimbeck, G. R. (1999) Acid rain impacts on calcium nutrition
53 and forest health. Bioscience 49: 789-800.
54 Del Monte, M.; Vittori, O. (1985) Air pollution and stone decay: the case of Venice. Endeavour 9: 117-122.
April 2002 4-216 DRAFT-DO NOT QUOTE OR CITE
-------
1 Diakumaku, E.; Gorbushina, A. A.; Krumbein, W. E.; Panina, L.; Soukharjevski, S. (1995) Black fungi in marble
2 and limestones~an aesthetical, chemical and physical problem for the conservation of monuments.
3 In: Saiz-Jimenez, C., ed. The deterioration of monuments: proceedings of the 2nd international symposium
4 onbiodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci. Total Environ. 167: 295-304.
5 Dickerson, R. R.; Kondragunta, S.; Stenchikov, G.; Civerolo, K. L.; Doddridge, B. G.; Holben, B. N. (1997)
6 The impact of aerosols on solar ultraviolet radiation and photochemical smog. Science (Washington, DC)
7 278: 827-830.
8 Dobbins, R. A.; Mulholland, G. W.; Bryner, N. P. (1994) Comparison of a fractal smoke optics model with light
9 extinction measurements. Atmos. Environ. 28: 889-897.
10 Dolske, D. A. (1988) Dry deposition of airborne sulfate and nitrate to soybeans. Environ. Pollut. 53: 1-12.
11 Dolske, D. A. (1995) Deposition of atmospheric pollutants to monuments, statues, and buildings. Sci. Total
12 Environ. 167: 15-31.
13 Dolske, D. A.; Gatz, D. F. (1984) Field intercomparison of sulfate dry deposition monitoring and measurement
14 methods: preliminary results. In: Hicks, B. B., ed. Deposition both wet and dry: [papers presented at the
15 American Chemical Society acid rain symposium]; March 1982; Las Vegas, NV. Boston, MA: Butterworth
16 Publishers; pp. 121-131. (Teasley, J. L, ed. Acid precipitation series: v. 4).
17 Draaijers, G. P. J.; Erisman, J. W. (1993) Atmospheric sulphur deposition to forest stands: throughfall estimates
18 compared to estimates from inference. Atmos. Environ. Part A 27: 43-55.
19 Draaijers, G. P. J.; Ivens, W. P. M. F.; Bleuten, W. (1988) Atmospheric deposition in forest edges measured by
20 monitoring canopy throughfall. Water Air Soil Pollut. 42: 129-136.
21 Draaijers, G. P. J.; vanEk, R.; Meijers, R. (1992) Research on the impact of forest stand structure on atmospheric
22 deposition. Environ. Pollut. 75: 243-249.
23 Dreisinger, B. R.; McGovern, P. C. (1970) Monitoring atmospheric sulphur dioxide and correlating its effects on
24 crops and forests in the Sudbury area. In: Linzon, S. N., ed. Impact of air pollution on vegetation conference;
25 April; Toronto, ON, Canada. Pittsburgh, PA: Air Pollution Control Association.
26 Driscoll, C. T.; Van Dreason, R. (1993) Seasonal and long-term temporal patterns in the chemistry of Adirondack
27 lakes. Water Air Soil Pollut. 67: 319-344.
28 Driscoll, C. T.; Wyskowski, B. J.; DeStaffan, P.; Newton, R. M. (1989) Chemistry and transfer of aluminum in a
29 forested watershed in the Adirondack region of New York, USA. In: Lewis, T. E., ed. Environmental
30 chemistry and toxicology of aluminum. Chelsea, MI: Lewis Publishers, Inc.; pp. 83-105.
31 Driscoll, C. T.; Newton, R. M.; Gubala, C. P.; Baker, J. P.; Christensen, S. W. (1991) Adirondack mountains.
32 In: Charles, D. F., ed. Acidic deposition and aquatic ecosystems: regional case studies. New York, NY:
33 Springer-Verlag; pp. 133-202.
34 Driscoll, C. T.; Lawrence, G. B.; Bulger, A. J.; Butler, T. J.; Cronan, C. S.; Eagar, C.; Lambert, K. F.; Likens, G. E.;
35 Stoddard, J. L.; Weathers, K. C. (2001) Acidic deposition in the northeastern United States: sources and
36 inputs, ecosystem effects, and management strategies. BioScience 51: 180-198.
37 Duan, B.; Fairall, C. W.; Thomson, D. W. (1988) Eddy correlation measurements of the dry deposition of particles
38 in wintertime. J. Appl. Meteorol. 27: 642-652.
39 Duyzer, J. H.; Bouman, A. M. M.; Van Aalst, R. M.; Diederen, H. S. M. A. (1987) Assessment of dry deposition
40 fluxes of NH3 and NH4+ over natural terrains. In: Asman, W. A. H.; Diederen, H. S. M. A., eds. Ammonia
41 and acidification: proceedings of a symposium of the European Association for the Science of Air Pollution
42 (EURASAP); April; Bilthoven, The Netherlands. European Association for the Science of Air Pollution;
43 pp. 97-106.
44 Eaton, J. S.; Likens, G. E.; Bormann, F. H. (1973) Throughfall and stemflow chemistry in a northern hardwood
45 forest. J.Ecol. 61:495-508.
46 Edgerton, E. S.; Lavery, T. F.; Boksleitner, R. P. (1992) Preliminary data from the USEPA dry deposition network:
47 1989. Environ. Pollut. 75: 145-156.
48 Edgerton-Warburton, L. M.; Allen, E. B. (2000) Shifts in arbuscular mycorrhizal communities along an
49 anthropogenic gradient nitrogen deposition gradient. Ecol. Appl. 10: 484-496.
50 Edney, E. O. (1989) Paint coatings: controlled field and chamber experiments. Research Triangle Park,
51 NC: U.S. Environmental Protection Agency, Atmospheric Research and Exposure Assessment Laboratory;
52 report no. EPA/600/3-89/032. Available from: NTIS, Springfield, VA; PB89-189849.
53 Edney, E. O.; Stiles, D. C.; Corse, E. W.; Wheeler, M. L.; Spence, J. W.; Haynie, F. H.; Wilson, W. E., Jr. (1988)
54 Field study to determine the impact of air pollutants on the corrosion of galvanized steel. Mater. Perform.
55 27:47-50.
April 2002 4-217 DRAFT-DO NOT QUOTE OR CITE
-------
1 Edney, E. O.; Cheek, S. F.; Corse, E. W.; Spence, J. W.; Haynie, F. H. (1989) Atmospheric weathering caused by
2 dry deposition of acidic compounds. J. Environ. Sci. Health Part A 24: 439-457.
3 Edwards, R. S.; Claxton, S. M. (1964) The distribution of air-borne salt of marine origin in the Aberystwyth area.
4 J.Appl.Ecol. 1:253-263.
5 El-Shobokshy, M. S. (1985) The dependence of airborne paniculate deposition on atmospheric stability and surface
6 conditions. Atmos. Environ. 19: 1191-1197.
7 Eldering, A.; Cass, G. R.; Moon, K. C. (1994) An air monitoring network using continuous particle size distribution
8 monitors: connecting pollutant properties to visibility via Mie scattering calculations. Atmos. Environ.
9 28:2733-2749.
10 Eldering, A.; Hall, J. R.; Hussey, K. J.; Cass, G. R. (1996) Visibility model based on satellite-generated landscape
11 data. Environ. Sci. Tech. 30: 361-370.
12 Eldred, R. A.; Cahill, T. A. (1994) Trends in elemental concentrations of fine particles at remote sites in the
13 United States of America. Atmos. Environ. 28: 1009-1019.
14 Eldred, R. A.; Cahill, T. A.; Feeney, P. J. (1993) Comparison of independent measurements of sulfur and sulfate in
15 the IMPROVE network. Presented at: 86th annual meeting and exhibition of the Air & Waste Management
16 Association; June; Denver, CO. Pittsburgh, PA: Air & Waste Management Association; paper no.
17 93-RA-l 10.02.
18 Eldred, R. A.; Cahill, T. A.; Flocchini, R. G. (1997) Composition of PM25 andPM10 aerosols in the IMPROVE
19 network. In: Special issue on aerosols and atmospheric optics-part I, [published papers from the international
20 specialty conference; September 1994; Snowbird, UT]. J. Air Waste Manage. Assoc. 47: 194-203.
21 Ellenberg, H. (1987) Floristic changes due to eutrophication. In: Asman, W. A. H.; Diederen, S. M. A., eds.
22 Ammonia and acidification: proceedings of a symposium of the European Association for the Science of Air
23 Pollution (EURASAP); April; Bilthoven, The Netherlands. European Association for the Science of Air
24 Pollution; pp. 301-308.
25 Eller, B. M. (1977) Road dust induced increase of leaf temperature. Environ. Pollut. 13: 99-107.
26 Eriksson, P.; Johansson, L.-G.; Strandberg, H. (1993) Initial stages of copper corrosion in humid air containing SO2
27 and NO2. J. Electrochem. Soc. 140: 53-59.
28 Erisman, J. W. (1993) Acid deposition onto nature areas in the Netherlands; part II. Throughfall measurements
29 compared to deposition estimates. Water Air Soil Pollut. 71: 81-99.
30 Erisman, J.-W.; Vermetten, A. W. M.; Asman, W. A. H.; Waijers-IJpelaan, A.; Slanina, J. (1988) Vertical
31 distribution of gases and aerosols: the behaviour of ammonia and related components in the lower
32 atmosphere. Atmos. Environ. 22: 1153-1160.
33 Erisman, J. W.; Van Pul, A.; Wyers, P. (1994) Parametrization of surface resistance for the quantification of
34 atmospheric deposition of acidifying pollutants and ozone. Atmos. Environ. 28: 2595-2607.
35 Erlick, C.; Frederick, J. E.; Saxena, V. K.; Wenny, B. N. (1998) Atmospheric transmission in the ultraviolet and
36 visible: aerosols in cloudy atmospheres. J. Geophys. Res. 103: 31,541-31,556.
37 Eshleman, K. N.; Miller-Marshall, L. M.; Webb, J. R. (1995) Long-term changes in episodic acidification of
38 streams in Shenandoah National Park, Virginia (U.S.A.). Water Air Soil Pollut. 85: 517-522.
39 Estupinan, J. B.; Raman, S.; Crescenti, G. H.; Streicher, J. J.; Barnard, W. F. (1996) Effects of clouds and haze on
40 UV-B radiation. J. Geophys. Res. 101: 16,807-16,816.
41 Evans, E. G.; Pitchford, M. (1991) Federal visibility monitoring in the eastern United States. Presented at:
42 84th annual meeting and exhibition of the Air & Waste Management Association; June; Vancouver, BC,
43 Canada. Pittsburgh, PA: Air & Waste Management Association; paper no. 91-49.6.
44 Fahey, T. J.; Yavitt, J. B.; Joyce, G. (1988) Precipitation and throughfall chemistry mPinus contorta ssp. latifolia
45 ecosystems, southeastern Wyoming. Can. J. For. Res. 18: 337-345.
46 Falkengren-Grerup, U. (1987) Long-term changes in pH of forest soils in southern Sweden. Environ. Pollut.
47 43:79-90.
48 Falkengren-Grerup, U. (1998) Nitrogen response of herbs and graminoids in experiments with simulated acid soil
49 solution. Environ. Pollut. 102(suppl. 1): 93-99.
50 Farber, R. J.; Welsing, P. R.; Rozzi, C. (1994) PM10 and ozone control strategy to improve visibility in the
51 Los Angeles Basin. Atmos. Environ. 28: 3277-3283.
52 Farmer, A. M. (1993) The effects of dust on vegetation—a review. Environ. Pollut. 79: 63-75.
53 Federal Register. (1999) Regional haze regulations; final rule. F. R. (July 1) 64: 35,713-35,774.
54 Fenn, M. E.; Keifer, J. W. (1999) Throughfall deposition of nitrogen and sulfur in a Jeffrey pine forest in the
55 San Gabriel Mountains, Southern California. Environ. Pollut. 104: 179-187.
April 2002 4-218 DRAFT-DO NOT QUOTE OR CITE
-------
1 Fenn, M. E.; Poth, M. A. (1999) Temporal and spatial trends in streamwater nitrate concentrations in the
2 San Bernardino Mountains, southern California. J. Environ. Qual. 28: 822-836.
3 Fenn, M. E.; Poth, M. A.; Johnson, D. W. (1996) Evidence for nitrogen saturation in the San Bernardino Mountains
4 in southern California. For. Ecol. Manage. 82: 211-230
5 Fenn, M. E.; Poth, M. A.; Aber, J. D.; Baron, J. S.; Bormann, B. T.; Johnson, D. W.; Lemly, A. D.; McNulty, S. G.;
6 Ryan, D. F.; Stottlemyer, R. (1998) Nitrogen excess in North American ecosystems: predisposing factors,
7 ecosystem responses, and management strategies. Ecol. Appl. 8: 706-733.
8 Fink, F. W.; Buttner, F. H.; Boyd, W. K. (1971) Final report on technical-economic evaluation of air-pollution
9 corrosion costs on metals in the U.S. Columbus, OH: Battelle Memorial Institute, Columbus Laboratories;
10 EPA contract no. CPA 70-86; EPA report no. APTD-0654. Available from: NTIS, Springfield, VA;
11 PB-198-453.
12 Flowers, E. C.; McCormick, R. A.; Kurfis, K. R. (1969) Atmospheric turbidity over the United States, 1961-1966.
13 J. Appl. Meteorol. 8: 955-962.
14 Flum, T.; Nodvin, S. C. (1995) Factors affecting streamwater chemistry in the Great Smoky Mountains, USA.
15 Water Air Soil Pollut. 85: 1707-1712.
16 Fochtman, E. G.; Langer, G. (1957) Automobile paint damaged by airborne iron particles. J. Air Pollut. Control
17 Assoc. 6: 243-247.
18 Foltescu, V. L.; Isakson, J.; Selin, E.; Stikans, M. (1994) Measured fluxes of sulphur, chlorine and some
19 anthropogenic metals to the Swedish west coast. Atmos. Environ. 28: 2639-2649.
20 Foster, J. R. (1991) Effects of organic chemicals in the atmosphere on terrestrial plants. In: Moser, T. J.; Barker,
21 J. R.; Tingey, D. T., eds. Ecological exposure and effects of airborne toxic chemicals: an overview. Corvallis,
22 OR: U.S. Environmental Protection Agency, Environmental Research Laboratory; pp. 60-89; EPA report no.
23 EPA-600/3-91-001. Available from: NTIS, Springfield, VA; PB91-148460.
24 Foster, N. W.; Nicolson, J. A.; Hazlett, P. W. (1989) Temporal variation in nitrate and nutrient cations in drainage
25 waters from a deciduous forest. J. Environ. Qual. 18: 238-244.
26 Fowler, D.; Cape, J. N.; Unsworth, M. H. (1989) Deposition of atmospheric pollutants on forests. Philos. Trans. R.
27 Soc. London B 324: 247-265.
28 Fowler, D.; Duyzer, J. H.; Baldocchi, D. D. (1991) Inputs of trace gases, particles and cloud droplets to terrestrial
29 surfaces. Proc. R. Soc. Edinburgh Sect. B: Biol. Sci. 97: 35-59.
30 Fox, D. G.; Malm, W. C.; Mitchell, B.; Fisher, R. W. (1999) Where there's fire, there's smoke: fine paniculate and
31 regional haze. EM November: 15-24.
32 Frederick, J. E.; Qu, Z.; Booth, C. R. (1998) Ultraviolet radiation at sites on the Antarctic coast. J. Photochem.
33 Photobiol. B 68: 183-190.
34 Freedman, B.; Hutchinson, T. C. (1981) Sources of metal and elemental contamination of terrestrial environments.
3 5 In: Lepp, N. W., ed. Effect of heavy metal pollution on plants. Volume 2: metals in the environment. London,
36 United Kingdom: Applied Science Publishers; pp. 35-94.
37 Freedman, D. M.; Zahm, S. H.; Dosemeci, M. (1997) Residential and occupational exposure to sunlight and
3 8 mortality from non-Hodgkin's lymphoma: composite (threefold) case-control study. Br. Med. J. BMJ
39 314: 1451-1455.
40 Freudenburg, W. R.; Alario, M. (1999) What ecologists can learn from nuclear scientists. Ecosystems 2: 286-291.
41 Fuller, K. A.; Malm, W. C.; Kreidenweis, S. M. (1999) Effects of mixing on extinction by carbonaceous particles.
42 J. Geophys. Res. 104: 15,941-15,954.
43 Ftirst, P.; Krause, G. H. M.; Hein, D.; Delschen, T.; Wilmers, K. (1993) PCDD/PCDF in cow's milk in relation to
44 their levels in grass and soil. Chemosphere 27: 1349-1357.
45 Gaggi, C.; Bacci, E.; Calamari, D.; Fanelli, R. (1985) Chlorinated hydrocarbons in plant foliage: an indication of the
46 tropospheric contamination level. Chemosphere 14: 1673-1686.
47 Gallagher, M. W.; Choularton, T. W.; Morse, A. P.; Fowler, D. (1988) Measurements of the size dependence of
48 cloud droplet deposition at a hill site. Q. J. R. Meteorol. Soc. 114: 1291-1303.
49 Garcia-Valles, M.; Vendrell-Saz, M.; Molera, J.; Blazquez, F. (1998) Interaction of rock and atmosphere: patinas on
50 Mediterranean monuments. Environ. Geol. 36(1-2): 137-149.
51 Garland, J. A. (1978) Dry and wet removal of sulphur from the atmosphere. Atmos. Environ. 12: 349-362.
52 Garland, C.; Garland, F. (1980) Do sunlight and vitamin D reduce the likelihood of colon cancer? Int. J. Epidemiol.
53 9:227-231.
54 Garland, F. C.; Garland, C. F.; Gorham, E. D.; Young, J. F. (1990) Geographic variation in breast cancer mortality
55 in the United States: a hypothesis involving exposure to solar radiation. Prev. Med. 19: 614-622.
April 2002 4-219 DRAFT-DO NOT QUOTE OR CITE
-------
1 Garner, J. H. B. (1991) Ozone exposure and nitrogen loading in the forests of eastern North America. In: Berglund,
2 R. L.; Lawson, D. R.; McKee, D. J., eds. Tropospheric ozone and the environment: papers from an
3 international conference; March 1990; Los Angeles, CA. Pittsburgh, PA: Air & Waste Management
4 Association; pp. 289-310. (A&WMA transactions series no. TR-19).
5 Garner, J. H. B. (1994) Nitrogen oxides, plant metabolism, and forest ecosystem response. In: Alscher, R. G.;
6 Wellburn, A. R., eds. Plant responses to the gaseous environment: molecular, metabolic and physiological
7 aspects, [3rd international symposium on air pollutants and plant metabolism]; June 1992; Blacksburg, VA.
8 London, United Kingdom: Chapman & Hall; pp. 301-314.
9 Garner, J. H. B.; Pagano, T.; Cowling, E. B. (1989) An evaluation of the role of ozone, acid deposition, and other
10 airborne pollutants in the forests of eastern North America. Asheville, NC: U.S. Department of Agriculture,
11 Forest Service, Southeastern Forest Experiment Station; general technical report SE-59.
12 Garten, C. T., Jr. (1988) Fate and distribution of sulfur-35 in yellow poplar and red maple trees. Oecologia
13 76:43-50.
14 Garten, C. T., Jr.; Hanson, P. J. (1990) Foliar retention of 15N-nitrate and 15N-ammoniumby red maple
15 (Acer rubrum) and white oak (Quercus alba) leaves from simulated rain. Environ. Exp. Bot. 30: 333-342.
16 Garten, C. T., Jr.; Bondietti, E. A.; Lomax, R. D. (1988) Contribution of foliar leaching and dry deposition to
17 sulfate in net throughfall below deciduous trees. Atmos. Environ. 22: 1425-1432.
18 Gawel, J. E.; Ahner, B. A.; Friedland, A. J.; Morel, F. M. M. (1996) Role for heavy metals in forest decline
19 indicated by phytochelatin measurements. Nature 3 81: 64-65.
20 Gazzi, M.; Georgiadis, T.; Vicentini, V. (2001) Distant contrast measurements through fog and thick haze.
21 Atmos. Environ. 35: 5143-5149.
22 Geron, C.; Rasmussen, R.; Arnts, R. R.; Guenther, A. (2000) A review and synthesis of monoterpene speciation
23 from forests in the United States. Atmos. Environ. 34: 1761-1781.
24 Ghedini, N.; Gobbi, G.; Sabbioni, C.; Zappia, G. (2000) Determination of elemental and organic carbon on
25 damaged stone monuments. Atmos. Environ. 34: 4383-4391.
26 Gildon, A.; Tinker, P. B. (1983) Interactions of vesicular-arbuscular mycorrhizal infection and heavy metals in
27 plants: I. the effects of heavy metals on the development of vesicular-arbuscular mycorrhizas. New Phytol.
28 95:247-261.
29 Gilliam, F. S.; Adams, M. B.; Yurish, B. M. (1996) Ecosystem nutrient responses to chronic nitrogen inputs at
30 Fernow Experimental Forest, West Virginia. Can. J. For. Res. 26: 196-205.
31 Gobbi, G.; Zappia, G.; Sabbioni, C. (1998) Sulphite quantification on damaged stones and mortars. Atmos. Environ.
32 32:783-789.
33 Goodwin, J. E.; Sage, W.; Tilly, G. P. (1969) Study of erosion by solid particles. Proc. Inst. Mech. Eng.
34 184:279-292.
35 Gorham, E. D.; Garland, C. F.; Garland, F. C. (1989) Acid haze air pollution and breast and colon cancer mortality
36 in 20 Canadian cities. Can. J. Public Health 80: 96-100.
37 Gorham, E. D.; Garland, F. C.; Garland, C. F. (1990) Sunlight and breast cancer incidence in the USSR. Int. J.
38 Epidemiol. 19: 820-824.
39 Graedel, T. E.; Nassau, K.; Franey, J. P. (1987) Copper patinas formed in the atmosphere—I. introduction.
40 Corros. Sci. 27: 639-657.
41 Grand Canyon Visibility Transport Commission. (1996) Report of the Grand Canyon Visibility Transport
42 Commission to the United States Environmental Protection Agency: recommendations for improving western
43 vistas. Denver, CO: Grand Canyon Visibility Transport Commission.
44 Grant, L. D. (1989) Health effects issues associated with regional and global air pollution problems. In: Proceedings
45 of the changing atmosphere: implications for global security; June, 1988; Toronto, Canada. Geneva,
46 Switzerland: World Meteorological Organization; report no. WMO-No. 710.
47 Grant, W. B. (2002) An estimate of premature cancer mortality in the U.S. due to inadequate doses of solar
48 ultraviolet-B radiation. Cancer 6: 1867-1875.
49 Grattan, S. R.; Maas, E. V.; Ogata, G. (1981) Foliar uptake and injury from saline aerosol. J. Environ. Qual.
50 10:406-409.
51 Gravenhorst, G.; Hoefken, K. D.; Georgii, H. W. (1983) Acidic input to a beech and spruce forest. In: Beilke, S.;
52 Elshout, A. J., eds. Acid deposition: proceedings of the CEC workshop organized as part of the concerted
53 action "Physico-chemical behaviour of atmospheric pollutants;" September 1982; Berlin, Federal Republic of
54 Germany. Dordrecht, The Netherlands: D. Reidel Publishing Company; pp. 155-171.
April 2002 4-220 DRAFT-DO NOT QUOTE OR CITE
-------
1 Great Lakes Regional Assessment Group (GLRAG). (2000) Preparing for a changing climate: the potential
2 consequences of climate variability and change. Great Lakes overview. Washington, DC: U.S. Environmental
3 Protection Agency; Office of Research and Development; U.S. Global Change Research Program
4 (USGCRP) and Ann Arbor, MI: University of Michigan; Atmospheric, Oceanic and Space Sciences
5 Department.
6 Greenberg, B. M.; Wilson, M. I.; Huang, X.-D.; Duxbury, C. L.; Gerhardt, K. E.; Gensemer, R. W. (1997)
7 The effects of ultraviolet-B radiation on higher plants. In: Wang, W.; Gorsuch, J. W.; Hughes, J. S., eds.
8 Plants for Environmental Studies. New York, NY: Lewis Publishers; pp. 1-35.
9 Grennfelt, P. (1987) Deposition processes for acidifying compounds. Environ. Technol. Lett. 8: 515-527.
10 Grimm, N. B.; Grove, J. M.; Pickett, S. T. A.; Redman, C. L. (2000) Integrated approaches to long-term studies of
11 urban ecological systems. Bioscience 50: 571-584.
12 Guderian, R. (1977) Accumulation of pollutants in plant organs. In: Air pollution: phytotoxicity of acidic gases and
13 its significance in air pollution control. Berlin, Germany: Springer-Verlag; pp. 66-74. (Ecological studies:
14 v. 22).
15 Guderian, R., ed. (1985) Air pollution by photochemical oxidants: formation, transport, control, and effects on
16 plants. New York, NY: Springer-Verlag. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.;
17 Remmert, H., eds. Ecological studies: analysis and synthesis: v. 52).
18 Guderian, R. (1986) Terrestrial ecosystems: paniculate deposition. In: Legge, A. H.; Krupa, S. V., eds.
19 Air pollutants and their effects on the terrestrial ecosystem. New York, NY: John Wiley & Sons;
20 pp. 339-363. (Advances in environmental science and technology: v. 18).
21 Gundersen, P.; Callesen, I.; De Vries, W. (1998a) Nitrate leaching in forest ecosystems is related to forest floor C/N
22 ratios. Environ. Pollut. 102(suppl. 1): 403-407.
23 Gundersen, P.; Emmett, B. A.; Kjonaas, O. J.; Koopmans, C. J.; Tietema, A. (1998b) Impact of nitrogen deposition
24 on nitrogen cycling in forests: a synthesis of NITREX data. For. Ecol. Manage. 101: 37-55.
25 Hamilton, R. S.; Revitt, D. M.; Vincent, K. J.; Butlin, R. N. (1995) Sulphur and nitrogen paniculate pollutant
26 deposition on to building surfaces. In: Saiz-Jimenez, C., ed. The deterioration of monuments: proceedings of
27 the 2nd international symposium on biodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci.
28 Total Environ. 167: 57-66.
29 Han, Q.; Rossow, W. B.; Lacis, A. A. (1994) Near-global survey of effective droplet radii in liquid water clouds
30 using ISCCP data. J. Clim. 7: 465-497.
31 Hanchette, C. L.; Schwartz, G. G. (1992) Geographic patterns of prostate cancer mortality. Cancer 70: 2861-2869.
32 Haneef, S. J.; Johnson, J. B.; Thompson, G. E.; Wood, G. C. (1993) Simulation of the degradation of limestones
33 and dolomitic sandstone under dry deposition conditions. In: Corrosion control for low-cost reliability:
34 preceedings [of the] 12th international corrosion congress. Volume 2: process industries plant operation.
3 5 Houston, TX: National Association of Corrosion Engineers; pp. 700-710.
36 Hansen, K.; Draaijers, G. P. J.; Ivens, W. P. M. F.; Gundersen, P.; van Leeuwen, N. F. M. (1994) Concentration
37 variations in rain and canopy throughfall collected sequentially during individual rain events. Atmos.
38 Environ. 28:3195-3205.
39 Hanson, P. J.; Lindberg, S. E. (1991) Dry deposition of reactive nitrogen compounds: a review of leaf, canopy and
40 non-foliar measurements. Atmos. Environ. Part A 25: 1615-1634.
41 Hartge, P.; Devessa, S. S.; Graumen, D.; Fears, T. R.; Fraumeni, J. F. (1996) Non-Hodgkin's lymphoma and
42 sunlight. J. Natl. Cancer Inst. 88: 298-300.
43 Harwell, M. A.; Myers, V; Young, T.; Bartuska, A.; Gassman, N.; Gentile, J. H.; Harwell, C. C.; Appelbaum, S.;
44 Barko, J.; Causey, B.; Johnson, C.; McLean, A.; Smola, R.; Templet, P.; Tosini, S. (1999) A framework for
45 an ecosystem integrity report card. BioScience 49: 543-556.
46 Hauhs, M. (1989) Lange Bramke: an ecosystem study of a forested catchment. In: Adriano, D. C.; Havas, M., eds.
47 Acidic precipitation: volume 1, case studies. New York, NY: Springer-Verlag; pp. 275-305. (Adriano, D. C.;
48 Salomons, W., eds. Advances in environmental science).
49 Haynie, F. H. (1976) Air pollution effects on stress induced intergranular corrosion of 7005-T53 aluminum alloy.
50 In: Stress corrosion-new approaches. Philadelphia, PA: American Society for Testing and Materials;
51 pp. 32-43. (ASTM special technical publication 610).
52 Haynie, F. H. (1980) Theoretical air pollution and climate effects on materials confirmed by zinc corrosion data.
53 In: Sereda, P. J.; Litvan, G. G., eds. Durability of building materials and components: proceedings of the
54 1st international conference; August 1978; Ottawa, ON, Canada. Philadelphia, PA: American Society for
55 Testing and Materials; pp. 157-175. (ASTM special technical publication 691).
April 2002 4-221 DRAFT-DO NOT QUOTE OR CITE
-------
1 Haynie, F. H. (1986) Environmental factors affecting corrosion of weathering steel. In: Baboian, R., ed. Materials
2 degradation caused by acid rain: developed from the 20th state-of-the-art symposium of the American
3 Chemical Society; June 1985; Arlington, VA. Washington, DC: American Chemical Society; pp. 163-171.
4 (ACS symposium series 318).
5 Haynie, F. H.; Lemmons, T. J. (1990) Paniculate matter soiling of exterior paints at a rural site. Aerosol Sci.
6 Technol. 13: 356-367.
7 Haynie, F. H.; Upham, J. B. (1974) Correlation between corrosion behavior of steel and atmospheric pollution data.
8 In: Coburn, S. K., ed. Corrosion in natural environments: presented at the 76th annual meeting American
9 Society for Testing and Materials; June 1973; Philadelphia, PA. Philadelphia, PA: American Society for
10 Testing and Materials; pp. 33-51. (ASTM special technical publication 558).
11 Haywood, J. M.; Shine, K. P. (1995) The effect of anthropogenic sulfate and soot aerosol on the clear sky planetary
12 radiation budget. Geophys. Res. Lett. 22: 603-606.
13 Haywood, J. M.; Ramaswamy, V.; Soden, B. J. (1999) Tropospheric aerosol climate forcing in clear-sky satellite
14 observations over the oceans. Science (Washington, DC) 283: 1299-1303.
15 Heal, G. (2000) Valuing ecosystem services. Ecosystems 3: 24-30.
16 Heath, J. A.; Huebert, B. J. (1999) Cloudwater deposition as a source of fixed nitrogen in a Hawaiian montane
17 forest. Biogeochemistry 44: 119-134.
18 Hedin, L. O.; Granat, L.; Likens, G. E.; Buishand, T. A.; Galloway, J. N.; Butler, T. J.; Rodhe, H. (1994)
19 Steep declines in atmospheric base cations in regions of Europe and North America. Nature (London)
20 367: 351-354.
21 Heichel, G. H.; Hankin, L. (1972) Particles containing lead, chlorine, and bromine detected on trees with an electron
22 microprobe. Environ. Sci. Technol. 6: 1121-1122.
23 Heichel, G. H.; Hankin, L. (1976) Roadside coniferous windbreaks as sinks for vehicular lead emissions. J. Air
24 Pollut. Control Assoc. 26: 767-770.
25 Heinsdorf, D. (1993) The role of nitrogen in declining Scots pine forests (Firms sylvestris) in the lowland of
26 East Germany. Water Air Soil Pollut. 69:21-35.
27 Hess, A.; Iyer, H.; Malm, W. (2001) Linear trend analysis: a comparison of methods. Atmos. Environ. 35:
28 5211-5222.
29 Hicks, B. B. (1986) Differences in wet and dry particle deposition parameters between North America and Europe.
30 In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P. J., eds. Aerosols: research, risk assessment and
31 control strategies; proceedings of the second U.S.-Dutch international symposium; May 1985; Williamsburg,
32 VA. Chelsea, MI: Lewis Publishers; pp. 973-982.
33 Hicks, B. B.; Matt, D. R.; McMillen, R. T.; Womack, J. D.; Shelter, R. E. (1984) Eddy fluxes of nitrogen oxides to
34 a deciduous forest in complex terrain. In: Samson, P. J., ed. Transactions: the meteorology of acid deposition,
35 an APCA specialty conference; October 1983; Hartford, CT. Pittsburgh, PA: Air Pollution Control
36 Association; pp. 189-201.
37 Hicks, B. B.; Baldocchi, D. D.; Meyers, T. P.; Hosker, R. P., Jr.; Matt, D. R. (1987) A preliminary multiple
3 8 resistance routine for deriving dry deposition velocities from measured quantities. Water Air Soil Pollut.
39 36:311-330.
40 Hicks, B. B.; Matt, D. R.; McMillen, R. T.; Womack, J. D.; Wesely, M. L.; Hart, R. L.; Cook, D. R.; Lindberg,
41 S. E.; de Pena, R. G.; Thomson, D. W. (1989) A field investigation of sulfate fluxes to a deciduous forest.
42 J. Geophys. Res. [Atmos.] 94: 13,003-13,011.
43 Hinckley, T.; Ford, D.; Segura, G.; Sprugel, D. (1992) Key processes from tree to stand level. In: Wall, G., ed.
44 Implications of climate change for Pacific Northwest forest management. Waterloo, ON, Canada: University
45 of Waterloo, Department of Geography; pp. 33-43. (Department of Geography publication series: occasional
46 paper no. 15).
47 Hofken, K. D.; Gravenhorst, G. (1982) Deposition of atmospheric aerosol particles to beech- and spruce forest.
48 In: Georgii, H.-W.; Pankrath, J., eds. Deposition of atmospheric pollutants: proceedings of a colloquium;
49 November 1981; Oberursel/Taunus, Federal Republic of Germany. Dordrecht, The Netherlands: D. Reidel
50 Publishing Company; pp. 191-194.
51 Hofken, K. D.; Meixner, F. X.; Ehhalt, D. H. (1983) Deposition of atmospheric trace constituents onto different
52 natural surfaces. In: Pruppacher, H. R.; Semonin, R. G.; Slinn, W. G. N., eds. Precipitation scavenging, dry
53 deposition, and resuspension: v. 2, dry deposition and resuspension: proceedings of the fourth international
54 conference; November-December 1982; Santa Monica, CA. New York, NY: Elsevier; pp. 825-835.
April 2002 4-222 DRAFT-DO NOT QUOTE OR CITE
-------
1 Hofstra, G.; Hall, R. (1971) Injury on roadside trees: leaf injury on pine and white cedar in relation to foliar levels
2 of sodium and chloride. Can. J. Bot. 49: 613-622.
3 Hogan, G. D.; Rennenberg, H.; Fink, S. (1998) Role and effect of sulfur in tree biology. In: Maynard, D. G., ed.
4 Sulfur in the environment. New York, NY: Marcel Dekker, Inc.; pp. 173-217. (Books in soils, plants, and the
5 environment: v. 67).
6 Holbrow, G. L. (1962) Atmospheric pollution: its measurement and some effects on paint. J. Oil Colour Chem.
7 Assoc. 45: 701-718.
8 Holland, D. M; Principe, P. P.; Sickles, J. E., II. (1999) Trends in atmospheric sulfur and nitrogen species in the
9 eastern United States for 1989-1995. Atmos. Environ. 33: 37-49.
10 Hollander, M.; Wolfe, D. A. (1973) Nonparametric statistical methods. New York: John Wiley & Sons.
11 Holling, C. S. (1986) The resilience of terrestrial ecosystems: local surprise and global change. In: Clark, W. C.;
12 Munn, R. E., eds. Sustainable development of the biosphere. Cambridge, UK: Cambridge University;
13 pp. 292-317.
14 Hooper, D. U.; Vitousek, P. M. (1997) The effects of plant composition and diversity on ecosystem processes.
15 Science 277: 1302-1305.
16 Hornung, M.; Langan, S. J. (1999) Nitrogen deposition, sources, impacts and responses in natural and semi-natural
17 ecosystems. In: Bangan, S. J., ed. Impact of nitrogen deposition on natural ecosystems and semi-natural
18 ecosystems. Dordrect, Netherlands: Kluwer Academic Publishers; pp. 1-14. [Environmental Pollution, no.3].
19 Horvath, H. (1993) Atmospheric light absorption-a review. Atmos. Environ. Part A 27: 293-317.
20 Hosker, R. P., Jr.; Lindberg, S. E. (1982) Review: atmospheric deposition and plant assimilation of gases and
21 particles. Atmos. Environ. 16: 889-910.
22 Houghton, J. T.; Jenkins, G. J.; Ephraums, J. J., eds. (1990) Climate change: the IPCC scientific assessment.
23 Cambridge, MA: Cambridge University Press; p. 55.
24 Hoyt, D. V. (1978) A model for the calculation of solar global insolation. Sol. Energy 21: 27-35.
25 Huebert, B. J.; Luke, W. T.; Delany, A. C.; Brost, R. A. (1988) Measurements of concentrations and dry surface
26 fluxes of atmospheric nitrates in the presence of ammonia. J. Geophys. Res. [Atmos.] 93: 7127-7136.
27 Hughes, M. K. (1981) Cycling of trace metals in ecosystems. In: Lepp, N. W., ed. Effect of heavy metal pollution
28 on plants. Volume 2: metals in the environment. London, United Kingdom: Applied Science Publishers;
29 pp. 95-118.
30 Hughes, J. B.; Daily, G. C.; Ehrlich, P. R. (1997) The loss of population diversity and why it matters. In: Raven,
31 P. H., ed. Nature and Human Society: the quest for a sustainable world. Washington, DC: National Academy
32 Press; pp. 71-83.
33 Humphreys, F. R.; Lambert, M. J.; Kelly, J. (1975) The occurrence of sulphur deficiency in forests. In: McLachlan,
34 K. D., ed. Sulphur in Australasian agriculture. Sydney, Australia: Sydney University Press; pp. 154-162.
3 5 Huntington, T.G. (2000) The potential for calcium depletion in forest ecosystems of southeastern United States:
36 review and analysis. Global Biogeochem. Cycles 14: 623-638.
37 Husar, R. B.; Wilson, W. E. (1993) Haze and sulfur emission trends in the eastern United States. Environ. Sci.
38 Technol. 27: 12-16.
39 Husar, R. B.; Holloway, J. M.; Patterson, D. E.; Wilson, W. E. (1981) Spatial and temporal pattern of eastern U.S.
40 haziness: a summary. Atmos. Environ. 15: 1919-1928.
41 Husar, R. B.; Elkins, J. B.; Wilson, W. E. (1994) U.S. visibility trends, 1960-1992, regional and national. Presented
42 at: 87th annual meeting of the Air & Waste Management Association; June; Cincinnati, OH. Pittsburgh, PA:
43 Air & Waste Management Association; paper no. 94-MP3.05.
44 Hutchinson, A. J.; Johnson, J. B.; Thompson, G. E.; Wood, G. C.; Sage, P. W.; Cooke, M. J. (1992) The role of
45 fly-ash paniculate material and oxide catalysts in stone degradation. Atmos. Environ. 26A: 2795-2803.
46 Huttunen, S. (1994) Effects of air pollutants on epicuticular wax structure. In: Percy, K. E.; Cape, J. N.; Jagels, R.;
47 Simpson, C. J., eds. Air pollutants and the leaf cuticle. New York, NY: Springer-Verlag; pp. 81-96.
48 (NATO ASI series).
49 Hyer, K. E.; Webb, J. R.; Eshleman, K. N. (1995) Episodic acidification of three streams in Shenandoah National
50 Park, Virginia, USA. Water Air Soil Pollut. 85: 523-528.
51 Ibrahim, M.; Barrie, L. A.; Fanaki, F. (1983) An experimental and theoretical investigation of the dry deposition of
52 particles to snow, pine trees and artificial collectors. Atmos. Environ. 17: 781-788.
53
54
April 2002 4-223 DRAFT-DO NOT QUOTE OR CITE
-------
1 IMPROVE: interagency monitoring of protected visual environments [database]. (1998) [Data on gross chemical
2 composition of PM2.5 particles in rural areas of the eastern and western United States as of 1998]. Fort
3 Collins, CO: National Park Service (NPS); Cooperative Institute for Research in the Atmosphere (CIRA).
4 Available at: http://vista.cira.colostate.edu/improve/ [2001, January 26].
5 Intergovernmental Panel on Climate Change (IPCC). (1995) Climate change 1994: radiative forcing of climate
6 change and an evaluation of the IPCC IS92 emission scenarios. Cambridge, United Kingdom: University
7 Press.
8 Intergovernmental Panel on Climate Change (IPCC). (1996) Climate change 1996: contribution of working group
9 I to the second assessment of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom:
10 Cambridge University Press; p. 572.
11 Intergovernmental Panel on Climate Change (IPCC). (1998) The regional impacts of climate change: an assessment
12 of vulnerability. Cambridge, United Kingdom: Cambridge University Press.
13 Intergovernmental Panel on Climate Change (IPCC). (200la) Climate change 2001: the scientific basis.
14 Contribution of working group I to the third assessment report of the Intergovernmental Panel on Climate
15 Change. Cambridge, United Kingdom: Cambridge University Press.
16 Intergovernmental Panel on Climate Change (IPCC). (200 Ib) Climate change 2001: impacts, adaptation, and
17 vulnerability. Contribution of working group II to the third assessment report of the Intergovernmental Panel
18 on Climate Change. Cambridge, United Kingdom: Cambridge University Press.
19 Irving, P. M, ed. (1991) Acidic deposition: state of science and technology: volumes I-IV. Washington, DC:
20 The U.S. National Acid Precipitation Assessment Program.
21 Iyer, H.; Patterson, P.; Malm, W. C. (2000) Trends in the extremes of sulfur concentration distributions. J. Air
22 Waste Manage. Assoc. 50: 802-808.
23 Jacobson, M. Z. (1998) Studying the effects of aerosols on vertical photolysis rate coefficient and temperature
24 profiles over an urban airshed. J. Geophys. Res. [Atmos.] 103: 10,593-10,604.
25 Jacobson, M. Z. (1999) Isolating nitrated and aromatic aerosols and nitrated aromatic gases as sources of ultraviolet
26 light absorption. J. Geophys. Res. 104 (D3): 3527-3542.
27 Janzen, J. (1980) Extinction of light by highly nonspherical strongly absorbing colloidal particles:
28 spectrophotometric determination of volume distributions for carbon blacks. Appl. Opt. 19: 2977-2985.
29 Jaworski, N. A.; Howarth, R. W.; Hetling, L. J. (1997) Atmospheric deposition of nitrogen oxides onto the
3 0 landscape contributes to coastal eutrophication in the northeast United States. Environ. Sci. Technol. 31:
31 1995-2004.
32 Jensen, V. (1974) Decomposition of angiosperm tree leaf litter. In: Dickinson, C. H.; Pugh, G. J. F., eds. Biology of
33 plant litter decomposition: v. I. London, United Kingdom: Academic Press; pp. 69-104.
34 John, W.; Wall, S. M.; Ondo, J. L. (1985) Dry acid deposition on materials and vegetation: concentrations in
3 5 ambient air. Berkeley, CA: California Department of Health Services, Air and Industrial Hygiene Laboratory;
36 report no. CA/DOH/AIHL/SP-34. Available from: NTIS, Springfield, VA; PB85-241206.
37 Johnson, D. W. (1984) Sulfur cycling in forests. Biogeochemistry 1: 29-43.
38 Johnson, D. W. (1992) Base cation distribution and cycling. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric
3 9 deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
40 Springer-Verlag, Inc.; pp. 275-333. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
41 Ecological studies, analysis and synthesis: v. 91).
42 Johnson, D. W.; Lindberg, S. E., eds. (1992a) Atmospheric deposition and forest nutrient cycling: a synthesis of the
43 integrated forest study. New York, NY: Springer-Verlag, Inc. (Billings, W. D.; Golley, F.; Lange, O. L.;
44 Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
45 Johnson, D. W.; Lindberg, S. E., eds. (1992b) Nitrogen chemistry, deposition, and cycling in forests. In: Johnson,
46 D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated
47 forest study. New York, NY: Springer-Verlag, Inc.; pp. 150-213. (Billings, W. D.; Golley, F.; Lange, O. L.;
48 Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
49 Johnson, D. W.; Mitchell, M. J. (1998) Responses of forest ecosystems to changing sulfur inputs. In: Maynard,
50 D. G., ed. Sulfur in the environment. New York, NY: Marcel Dekker, Inc.; pp. 219-262. [Books in Soils,
51 Plants, and the Environment, v. 67].
52 Johnson, D. W.; Taylor, G. E. (1989) Role of air pollution in forest decline in eastern North America. Water Air
53 Soil Pollut. 48:21-43.
54 Johnson, D. W.; Todd, D. E., Jr. (1998) Harvesting effects on long-term changes in nutrient pools of mixed oak
55 forest. Soil Sci. Soc. Amer. J. 62: 1725-1735.
April 2002 4-224 DRAFT-DO NOT QUOTE OR CITE
-------
1 Johnson, D. W.; Henderson, G. S.; Huff, D. D.; Lindberg, S. E.; Richter, D. D.; Shriner, D. S.; Todd, D. E.;
2 Turner, J. (1982a) Cycling of organic and inorganic sulphur in a chestnut oak forest. Oecologia 54: 141-148.
3 Johnson, D. W.; Cole, D. W.; Bledsoe, C. S.; Cromack, K.; Edmonds, R. L.; Gessel, S. P.; Grier, C. C; Richards,
4 B. N.; Vogt, K. A. (1982b) Nutrient cycling in forests of the Pacific Northwest. In: Edmonds, R. L., ed.
5 Analysis of coniferous forest ecosystems in the Western United States. Stroudsburg, PA: Hutchinson Ross
6 Publishing Company; pp. 186-232. (US/IBP synthesis series: v. 14).
7 Johnson, D. W.; Van Miegroet, H.; Lindberg, S. E.; Todd, D. E.; Harrison, R. B. (1991a) Nutrient cycling in red
8 spruce forests of the Great Smoky Mountains. Can. J. For. Res. 21: 769-787.
9 Johnson, D. W.; Cresser, M. S.; Nilsson, S. L; Turner, J.; Ulrich, B.; Binkley, D.; Cole, D. W. (1991b) Soil changes
10 in forest ecosystems: evidence for and probable causes. Proc. R. Soc. Edinburgh Sect. B: Biol. Sci.
11 97B: 81-116.
12 Johnson, D. W.; Swank, W. T.; Vose, J. M. (1993) Simulated effects of atmospheric sulfur deposition on nutrient
13 cycling in a mixed deciduous forest. Biogeochemistry 23: 169-196.
14 Johnson, A. H.; Friedland, A. J.; Miller, E. K.; Siccama, T. G. (1994) Acid rain and soils of the Adirondacks. III.
15 Rates of soil acidification in a montane spruce-fir forest at Whiteface Mountain, New York. Can. J. For. Res.
16 24:663-669.
17 Johnson, D.W.; Susfalk, R.B.; Brewer, P.,F.; Swank, W.T. (1999) Simulated effects of reduced sulfur, nitrogen, and
18 base cation deposition on soils and solutions in southern Appalachian forests. J. Environ. Qual
19 28: 1336-1346.
20 Joslin, J. D.; Wolfe, M. H. (1992) Red spruce soil chemistry and root distribution across a cloud water deposition
21 gradient. Can. J. For. Res. 22: 893-904.
22 Joslin, J. D.; Wolfe, M. H. (1994) Foliar deficiencies of mature Southern Appalachian red spruce determined from
23 fertilizer trials. Soil Sci. Soc. Am. J. 58: 1572-1579.
24 Joslin, J. D.; Kelly, J. M.; Van Miegroet, H. (1992) Soil chemistry and nutrition of North American spruce-fir
25 stands: evidence for recent change. J. Environ. Qual. 21: 12-30.
26 Justus, C. G.; Murphey, B. B. (1994) Temporal trends in surface irradiance at ultraviolet wavelengths. J. Geophys.
27 Res. 99 (Dl): 1389-1394.
28 Kaminski, M. D.; Landsberger, S. (2000a) Heavy metals in urban soils of East St. Louis, IL, part I: total
29 concentration of heavy metals in soils. J. Air Waste Manage. Assoc. 50: 1667-1679.
30 Kaminski, M. D.; Landsberger, S. (2000b) Heavy metals in urban soils of East St. Louis, IL, part II: leaching
31 characteristics and modeling. J. Air Waste Manage. Assoc. 50: 1680-1687.
32 Karmoker, J. L.; Clarkson, D. T.; Saker, L. R.; Rooney, J. M.; Purves, J. V. (1991) Sulphate deprivation depresses
33 the transport of nitrogen to the xylem and the hydraulic conductivity of barley (Hordeum vulgare L.) roots.
34 Planta 185: 269-278.
35 Kaupp, H.; McLachlan, M. S. (1999) Atmospheric particle size distributions of polychlorinated debenzo-p-dioxins
36 and dibenzofurans (PCDD/Fs) and polycyclic aromatic hydrocarbons (PAHs) and their implications for wet
37 and dry deposition. Atmos. Environ. 33: 85-95.
38 Kelly, J.; Lambert, M. J. (1972) The relationship between sulphur and nitrogen in the foliage ofPinus radiata.
39 Plant Soil 37: 395-407.
40 Kelly, J. M.; Meagher, J. F. (1986) Nitrogen input/output relationships for three forested watersheds in eastern
41 Tennessee. In: Correll, D. L., ed. Watershed research perspectives. Washington, DC: Smithsonian Institution
42 Press; pp. 360-391.
43 Kenk, G.; Fischer, H. (1988) Evidence from nitrogen fertilisation in the forests of Germany. Environ. Pollut.
44 54: 199-218.
45 Kiehl, J. T.; Briegleb, B. P. (1993) The relative roles of sulfate aerosols and greenhouse gases in climate forcing.
46 Science (Washington, DC) 260: 311-314.
47 Klepper, B.; Craig, D. K. (1975) Deposition of airborne particulates onto plant leaves. J. Environ. Qual. 4: 495-499.
48 Kodama, Y.; Lee, S. D., eds. (1994) The 13th UOEH international symposium and the 2nd pan Pacific cooperative
49 symposium on impact of increased UV-B exposure on human health and ecosystem; October 1993;
50 Kitakyushu, Japan. Kitakyushu, Japan: University of Occupational and Environmental Health.
51 Koslow, E. E.; Smith, W. H.; Staskawicz, B. J. (1977) Lead-containing particles on urban leaf surfaces.
52 Environ. Sci. Technol. 11: 1019-1021.
53 Krajickova, A.; Mejstfik, V. (1984) The effect of fly-ash particles on the plugging of stomata. Environ. Pollut.
54 36:83-93.
April 2002 4-225 DRAFT-DO NOT QUOTE OR CITE
-------
1 Kripke, M. L. (1989) Health effects of stratospheric ozone depletion: an overview. In: Schneider, T.; Lee, S. D.;
2 Wolters, G. J. R.; Grant, L. D. eds. Atmospheric ozone research and its policy implications: proceedings of
3 the Third US-Dutch International Symposium; May, 1988; Nijmegen, The Netherlands. Amsterdam, The
4 Netherlands: Elsevier, pp. 795-802. [Studies in Environmental Science 35].
5 Krupa, S. V.; Legge, A. H. (1986) Single or joint effects of coarse and fine paniculate sulfur aerosols and ozone on
6 vegetation. In: Lee, S. D.; Schneider, T.; Grant, L. D.; Verkerk, P. J., eds. Aerosols: research, risk assessment
7 and control strategies, proceedings of the second U.S.-Dutch international symposium; May 1985;
8 Williamsburg, VA. Chelsea, MI: Lewis Publishers, Inc.; pp. 879-887.
9 Krupa, S. V.; Legge, A. H. (1998) Sulphur dioxide, paniculate sulphur and its impacts on a boreal forest ecosystem.
10 In: Ambasht, R. S., ed. Modern trends in ecology and environment. Leiden, The Netherlands: Backhuys
11 Publishers; pp. 285-306.
12 Kylin, H.; Grimvall, E.; Ostman, C. (1994) Environmental monitoring of polychlorinated biphenyls using pine
13 needles as passive samplers. Environ. Sci. Technol. 28: 1320-1324.
14 Lamaud, E.; Chapuis, A.; Fontan, J.; Serie, E. (1994) Measurements and parameterization of aerosol dry deposition
15 in a semi-arid area. Atmos. Environ. 28: 2461 -2471.
16 Lamersdorf, N. P.; Meyer, M. (1993) Nutrient cycling and acidification of a northwest German forest site with high
17 atmospheric nitrogen deposition. For. Ecol. Manage. 62: 323-354.
18 Laurenroth, W. K.; Milchunas, D. G. (1985) SO2 effects on plant community function. In: Winner, W. E.; Mooney,
19 H. A.; Goldstein, R. A., eds. Sulfur dioxide and vegetation: physiology, ecology, and policy issues. Stanford,
20 CA: Stanford University Press; pp. 454-477.
21 Lawrence, G. B.; Huntington, T. G. (1999) Soil-Calcium depletion linked to acid rain and forest growth in the
22 eastern United States. Reston, VA: U.S. Geological Survey. Available: http://btdqs.usgs.gov/acidrain [1999,
23 November 23].
24 Lefkowitz, E. S.; Garland, C. F. (1994) Sunlight, vitamin D, and ovarian cancer mortality rates in U.S. women. Int.
25 J. Epidemiol. 23: 1133-1136.
26 Legge, A. H. (1980) Primary productivity, sulfur dioxide, and the forest ecosystem: an overview of a case study.
27 In: Miller, P. R., ed. Proceedings of symposium on effects of air pollutants on Mediterranean and temperate
28 forest ecosystems; June; Riverside, CA. Berkeley, CA: U.S. Department of Agriculture, Pacific Southwest
29 Forest and Range Experiment Station; pp. 51-62; Forest Service general technical report no. PSW-43.
30 Available from: NTIS, Springfield, VA; PB81-133720.
31 Lerman, S. L.; Darley, E. F. (1975) Particulates. In: Mudd, J. B.; Kozlowski, T. T., eds. Responses of plants to air
32 pollution. New York, NY: Academic Press; pp. 141-158.
33 Leslie, M. (2001) Having a bad air day. Science 294: 2255.
34 Levin, S. A. (1998) Ecosystems and the biosphere as complex adaptive systems. Ecosystems 1: 431-436.
3 5 Li-Jones, X.; Prospero, J. M. (1998) Variations in the size distribution of non-sea-salt sulfate aerosol in the marine
36 boundary layer at Barbados: impact of African dust. J. Geophys. Res. 103: 16,073-16,084.
37 Likens, G. E.; Driscoll, C. T.; Buso, D. C. (1996) Long-term effects of acid rain: response and recovery of a forest
38 ecosystem. Science (Washington, DC) 272: 244-246.
39 Likens, G. E.; Driscoll, C. T.; Buso, D. C.; Siccama, T. G.; Johnson, C. E.; Lovett, G. M.; Fahey, T. J.; Reiners,
40 W. A.; Ryan, D. F.; Martin, C. W.; Bailey, S. W. (1998) The biogeochemistry of calcium at Hubbard Brook.
41 Biogeochemistry 41: 89-173.
42 Likens, G. E.; Butler, T. J.; Buso, D. C. (2001) Long- and short-term changes in sulfate deposition: effects of the
43 1990 Clean Air Act Amendments. Biogeochemistry 52: 1-11.
44 Lin, J.-M.; Fang, G.-C.; Holsen, T. M.; Noll, K. E. (1993) A comparison of dry deposition modeled from size
45 distribution data and measured with a smooth surface for total particle mass, lead and calcium in Chicago.
46 Atmos. Environ. Part A 27: 1131-1138.
47 Lindberg, S. E. (1992) Atmospheric deposition and canopy interactions of sulfur. In: Johnson, D. W.; Lindberg,
48 S. E., eds. Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated forest study.
49 New York, NY: Springer-Verlag, Inc.; pp. 74-90. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.;
50 Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
51 Lindberg, S. E.; Garten, C. T., Jr. (1988) Sources of sulphur in forest canopy throughfall. Nature (London)
52 336: 148-151.
53 Lindberg, S. E.; Harriss, R. C. (1981) The role of atmospheric deposition in an eastern U.S. deciduous forest.
54 Water Air Soil Pollut. 16: 13-31.
April 2002 4-226 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lindberg, S. E.; Lovett, G. M. (1985) Field measurements of particle dry deposition rates to foliage and inert
2 surfaces in a forest canopy. Environ. Sci. Technol. 19: 238-244.
3 Lindberg, S. E.; Lovett, G. M. (1992) Deposition and forest canopy interactions of airborne sulfur: results from the
4 integrated forest study. Atmos. Environ. Part A 26: 1477-1492.
5 Lindberg, S. E.; McLaughlin, S. B. (1986) Air pollutant interactions with vegetation: research needs in data
6 acquisition and interpretation. In: Legge, A. H.; Krupa, S. V., eds. Air pollutants and their effects on the
7 terrestrial ecosystem. New York, NY: John Wiley & Sons, Inc.; pp. 449-503. (Nriagu, J. O., ed. Advances in
8 environmental science and technology: v. 18).
9 Lindberg, S. E.; Owens, J. G. (1993) Throughfall studies of deposition to forest edges and gaps in montane
10 ecosystems. Biogeochemistry 19: 173-194.
11 Lindberg, S. E.; Harriss, R. C.; Turner, R. R. (1982) Atmospheric deposition of metals to forest vegetation. Science
12 (Washington, DC) 215: 1609-1611.
13 Lindberg, S. E.; Lovett, G. M.; Richter, D. D.; Johnson, D. W. (1986) Atmospheric deposition and canopy
14 interactions of major ions in a forest. Science (Washington, DC) 231: 141-145.
15 Lindberg, S. E.; Lovett, G. M.; Meiwes, K.-J. (1987) Deposition and forest canopy interactions of airborne nitrate.
16 In: Hutchinson, T. C.; Meema, K. M., eds. Effects of atmospheric pollutants on forests, wetlands and
17 agricultural ecosystems. Berlin, Federal Republic of Germany: Springer-Verlag; pp. 117-130. (NATO as
18 1 series: v. G16).
19 Lindberg, S. E.; Lovett, G. M.; Schaefer, D. A.; Bredemeier, M. (1988) Dry deposition velocities and
20 surface-to-canopy scaling factors for aerosol calcium from forest canopy throughfall. J. Aerosol Sci.
21 19: 1187-1190.
22 Lindberg, S. E.; Bredemeier, M.; Schaefer, D. A.; Qi, L. (1990) Atmospheric concentrations and deposition of
23 nitrogen and major ions in conifer forests in the United States and Federal Republic of Germany. Atmos.
24 Environ. Part A 24: 2207-2220.
25 Liousse, C.; Cachier, H.; Jennings, S. G. (1993) Optical and thermal measurements of black carbon aerosol content
26 in different environments: variation of the specific attenuation cross-section, sigma (a). Atmos. Environ.
27 Part A 27: 1203-1211.
28 Little, P. (1977) Deposition of 2.75, 5.0 and 8.5 um particles to plant and soil surfaces. Environ. Pollut.
29 12:293-305.
30 Little, P.; Wiffen, R. D. (1977) Emission and deposition of petrol engine exhaust Pb-I. deposition of exhaust Pb to
31 plant and soil surfaces. Atmos. Environ. 11: 437-447.
32 Liu, S. C.; McKeen, S. A.; Madronich, S. (1991) Effect of anthropogenic aerosols on biologically active ultraviolet
33 radiation. Geophys. Res. Lett. 18: 2265-2268.
34 Lokke, H.; Bak, J.; Falkengren-Grerup, U.; Finlay, R. D.; Ilvesniemi, H.; Nygaard, P. H.; Starr, M. (1996)
35 Critical loads of acidic deposition for forest soils: is the current approach adequate. Ambio 25: 510-516.
36 Lorente, J.; Redano, A.; de Cabo, X. (1994) Influence of urban aerosol on spectral solar irradiance. J. Appl.
37 Meteorol. 33:406-415.
38 Lorusso, S.; Marabelli, M.; Troili, M. (1997) Air pollution and the deterioration of historic monuments. J. Environ.
39 Pathol. Toxicol. Oncol. 16: 171-173.
40 Lovett, G. M. (1984) Atmospheric deposition to forests. In: Breece, L.; Hasbrouck, S., eds. Forest responses to
41 acidic deposition: proceedings of the U. S.-Canadian conference; August 1983; Orono, ME. Orono, ME:
42 University of Maine, Orono, Land and Water Resources Center; pp. 7-18.
43 Lovett, G. M. (1992) Atmospheric deposition and canopy interactions of nitrogen. In: Johnson, D. W.; Lindberg,
44 S. E., eds. Atmospheric deposition and forest nutrient cycling: a synthesis of the integrated forest study.
45 New York, NY: Springer-Verlag, Inc.; pp. 152-166. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.;
46 Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
47 Lovett, G. M. (1994) Atmospheric deposition of nutrients and pollutants in North America: an ecological
48 perspective. Ecol. Appl. 4: 629-650.
49 Lovett, G. M.; Lindberg, S. E. (1984) Dry deposition and canopy exchange in a mixed oak forest as determined by
50 analysis of throughfall. J. Appl. Ecol. 21: 1013-1027.
51 Lovett, G. M.; Lindberg, S. E. (1986) Dry deposition of nitrate to a deciduous forest. Biogeochemistry 2: 137-148.
52 Lovett, G. M.; Lindberg, S. E. (1992) Concentration and deposition of particles and vapors in a vertical profile
53 through a forest canopy. Atmos. Environ. Part A 26: 1469-1476.
54 Lovett, G. M.; Lindberg, S. E. (1993) Atmospheric deposition and canopy interactions of nitrogen in forests. Can.
55 J. For. Res. 23: 1603-1616.
April 2002 4-227 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lovett, G. M; Traynor, M. M; Pouyat, R. V.; Carreiro, M. M; Zhu, W.-X.; Baxter, J. W. (2000) Atmospheric
2 deposition to oak forest along an urban-rural gradient. Environ. Sci. Technol. 34: 4294-4300.
3 Lowenthal, D. H.; Rogers, C. F.; Saxena, P.; Watson, J. G.; Chow, J. C. (1995) Sensitivity of estimated light
4 extinction coefficients to model assumptions and measurement errors. Atmos. Environ. 29: 751-766.
5 Madronich, S. (1987) Photodissociation in the Atmosphere. 1. Actinic flux and the effects of ground reflections and
6 clouds. J. Geophys. Res. [Atmos.] 92: 9740-9752.
7 Malm, W. C. (1989) Atmospheric haze: its sources and effects on visibility in rural areas of the continental United
8 States. Environ. Monit. Assess. 12: 203-225.
9 Malm, W. C. (1999) Introduction to visibility. Fort Collins, CO: Colorado State University; Cooperative Institute
10 for Research in the Atmosphere (CIRA); NFS Visibility Program. Available:
11 http://www.epa.gov/air/visibility/introvis.pdf [2 April 2002].
12 Malm, W. C. (2000) Spatial and seasonal patterns and temporal variability of haze and its constituents in the United
13 States. Report III. Fort Collins, CO: Cooperative Institute for Research in the Atmosphere, Colorado State
14 University. Available: http://vista.cira.colostate.edu/improve/Publications/Reports/2000/2000.htm [22 March,
15 2002].
16 Malm, W. C.; Kreidenweis, S. M. (1997) The effects of models of aerosol hygroscopicity on the apportionment of
17 extinction. Atmos. Environ. 31: 1965-1976.
18 Malm, W. C.; Sisler, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
19 concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99: 1347-1370.
20 Malm, W. C.; Molenar, J. V.; Eldred, R. A.; Sisler, J. F. (1996) Examining the relationship among atmospheric
21 aerosols and light scattering and extinction in the Grand Canyon area. J. Geophys. Res. [Atmos.]
22 101: 19,251-19,265.
23 Malm, W. C.; Day, D.; Kreidenweis, S. M. (1997) Comparison of measured and reconstructed scattering during an
24 intensive field study at Great Smoky Mountains National Park. Presented at: 90th annual meeting and
25 exhibition of the Air & Waste Management Association; June; Toronto, ON, CA. Pittsburgh, PA: Air &
26 Waste Management Association; 97-WA70.02.
27 Malm, W. C.; Day, D. E.; Kreidenweis, S. M. (2000) Light scattering characteristics of aerosols as a function of
28 relative humidity: part I—a comparison of measured scattering and aerosol concentrations using the
29 theoretical models. J. Air Waste Manage. Assoc. 50: 686-700.
30 Malm, W. C.; Day, D. E.; Kreidenweis, S. M. (2000) Light scattering characteristics of aerosols as a function of
31 relative humidity: Part Il-a comparison of measured scattering and aerosol concentrations using the statistical
32 models. J. Air Waste Manage. Assoc. 50: 701-709.
33 Mansfeld, F. (1980) Regional air pollution study: effects of airborne sulfur pollutants on materials. Research
34 Triangle Park, NC: U.S. Environmental Protection Agency, Environmental Sciences Research Laboratory;
35 EPA report no. EPA-600/4-80-007. Available from: NTIS, Springfield, VA; PB81-126351.
36 Marchwiriska, E.; Kucharski, R. (1987) The combined influence of SO2 and heavy metal-containing particulates on
37 beans, carrots and parsley. Environ. Monit. Assess. 8: 11-25.
38 Markewitz, D.; Richter, D.D.; Allen, L. H.; Urrego, J. B. (1998) Three decades of observed soil acidification in the
39 Calhoun Experimental Forest: Has acid rain made a difference? Soil Sci. Soc. Amer. J. 62: 1428-1439.
40 Marschner, H. (1995) Sulfur. In: Mineral nutrition of higher plants. 2nd ed. London, United Kingdom: Academic
41 Press; pp. 255-265.
42 Martin, M. H.; Coughtrey, P. J. (1981) Impact of metals on ecosystem function and productivity. In: Lepp, N. W.,
43 ed. Effect of heavy metal pollution on plants: volume 2, metals in the environment. Barking, United
44 Kingdom: Applied Science Publishers; pp. 119-158. (Mellanby, K., ed. Pollution monitoring series).
45 Martin, C. E.; Gravatt, D. A.; Loeschen, V. S. (1992) Photo synthetic responses of three species to acute exposures
46 of nitrate- and sulphate-containing aerosols. Atmos. Environ. Part A 26: 381-391.
47 Mathai, C. V. (1995) The Grand Canyon visibility. EM (December): 20-31.
48 Matson, P. A.; Parton, W. J.; Power, A. G.; Swift, M. J. (1997) Agricultural intensification and ecosystem
49 properties. Science (Washington, DC) 277: 504-509.
50 May, P. F.; Till, A. R.; Gumming, M. J. (1972) Systems analysis of 35sulphur kinetics in pastures grazed by sheep.
51 J. Appl. Ecol. 9: 25-49.
52 McCartney, H. A.; Aylor, D. E. (1987) Relative contributions of sedimentation and impaction to deposition of
53 particles in a crop canopy. Agric. For. Meteorol. 40: 343-358.
April 2002 4-228 DRAFT-DO NOT QUOTE OR CITE
-------
1 McCune, D. C.; Silberman, D. H.; Mandl, R. H.; Weinstein, L. H.; Freudenthal, P. C.; Giardina, P. A. (1977)
2 Studies on the effects of saline aerosols of cooling tower origin on plants. J. Air Pollut. Control Assoc.
3 27:319-324.
4 McDonald, R. L.; Unni, C. K.; Duce, R. A. (1982) Estimation of atmospheric sea salt dry deposition: wind speed
5 and particle size dependence. J. Geophys. Res. C: Oceans Atmos. 87: 1246-1250.
6 McDonnell, M. J.; Pickett, S. T. A.; Groffman, P.; Bohlen, P.; Pouyat, R. V.; Zipperer, W. C.; Parmelee, R. W.;
7 Carreiro, M. M.; Medley, K. (1997) Ecosystem processes along an urban-to-rural gradient. Urban Ecosyst.
8 1:21-36.
9 McDow, S. R.; Vartiainen, M.; Sun, Q.; Hong, Y.; Yao, Y.; Kamens, R. M. (1995) Combustion aerosol water
10 content and its effect onpolycyclic aromatic hydrocarbon reactivity. Atmos. Environ. 29: 791-797'.
11 McGee, E. S.; Mossotti, V. G. (1992) Gypsum accumulation on carbonate stone. Atmos. Environ. Part B
12 26:249-253.
13 McGowan, T. F.; Lipinski, G. E.; Santoleri, J. J. (1993) New rules affect the handling of waste fuels. Chem. Eng.
14 (March): 122-128.
15 Mclntyre, N. E.; Knowles-Yanez, K.; Hope, D. (2000) Urban ecology as an indisciplinary field: differences in the
16 use of "urban" between the social and natural sciences. Urban Ecosyst. 4: 5-24.
17 McKay, W. A.; Garland, J. A.; Livesley, D.; Halliwell, C. M.; Walker, M. I. (1994) The characteristics of the
18 shore-line sea spray aerosol and the landward transfer of radionuclides discharged to coastal sea water.
19 Atmos. Environ. 28: 3299-3309.
20 McLachlan, M. S. (1996a) Bioaccumulation of hydrophobic chemicals in agricultural food chains. Environ. Sci.
21 Tech. 30: 252-259.
22 McLachlan, M. S. (1996b) Biological uptake and transfer of polychlorinated dibenzo-p-dioxins and dibenzofurans.
23 In: Hester, R. E.; Harrison, R. M., eds. Chlorinated organic micropollutants. Cambridge, United Kingdom:
24 Royal Society of Chemistry; pp. 31-52. (Issues in environmental science and technology).
25 McLachlan, M. S. (1999) Framework for the interpretation of measurements of SOCs in plants. Environ. Sci.
26 Technol. 33: 1799-1804.
27 McMahon, T. A.; Denison, P. J. (1979) Empirical atmospheric deposition parameters~a survey. Atmos. Environ.
28 13:571-585.
29 McMillen, R. T. (1988) An eddy correlation technique with extended applicability to non-simple terrain. Boundary
30 LayerMeteorol. 43:231-245.
31 McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
32 aerosols. Atmos. Environ. 23: 497-507.
33 McMurry, P. H.; Zhang, X.; Lee, C.-T. (1996) Issues in aerosol measurement for optics assessments. J. Geophys.
34 Res. [Atmos.] 101: 19,189-19,197.
35 Meiwes, K. J.; Khanna, P. K. (1981) Distribution and cycling of sulphur in the vegetation of two forest ecosystems
36 in an acid rain environment. Plant Soil 60: 369-375.
37 Mid-Atlantic Regional Assessment Team (MARAT). (2000) Preparing for a changing climate: the potential
3 8 consequences of climate variability and change. Mid-Atlantic overview. Washington, DC: U. S.
39 Environmental Protection Agency; Office of Research and Development; U.S. Global Change Research
40 Program (USGCRP) and University Park, PA: Pennsylvania State University.
41 Middleton, W. E. K. (1952) Vision through the atmosphere. Toronto, ON, Canada: University of Toronto Press.
42 Milford, J. B.; Davidson, C. I. (1987) The sizes of paniculate sulfate and nitrate in the atmosphere~a review.
43 JAPCA 37: 125-134.
44 Millar, C. S. (1974) Decomposition of coniferous leaf litter. In: Dickinson, C. H.; Pugh, G. J. F., eds. Biology of
45 plant litter decomposition: volume I. New York, NY: Academic Press; pp. 105-128.
46 Miller, P. R.; McBride, J. R., eds. (1999) Oxidant air pollution impacts in the Montane forests of southern
47 California: a case study of the San Bernardino Mountains. New York, NY: Springer-Verlag.
48 (Ecological Studies: v. 134.)
49 Miller, W. C.; Fornes, R. E.; Gilbert, R. D.; Speer, A.; Spence, J. (1992) Removal of CaCO3 extender in residential
50 coatings by atmospheric acidic deposition. In: Measurement of toxic and related air pollutants: proceedings
51 of the 1992 U.S. EPA/A&WMA international symposium. Pittsburgh, PA: Air & Waste Management
52 Association; pp. 129-134. (A&WMA publication VIP-25).
53 Miller, E. K.; Panek, J. A.; Friedland, A. J.; Kadlecek, J.; Mohnen, V. A. (1993) Atmospheric deposition to a
54 high-elevation forest at Whiteface Mountain, New York, USA. Tellus Ser. B 45: 209-227.
April 2002 4-229 DRAFT-DO NOT QUOTE OR CITE
-------
1 Mims, F. M., III. (1996) Significant reduction of UVB caused by smoke from biomass burning in Brazil.
2 Photochem. Photobiol. 64: 814-816.
3 Mitchell, M. I; Harrison, R. B.; Fitzgerald, J. W.; Johnson, D. W.; Lindberg, S. E.; Zhang, Y.; Autry, A. (1992a)
4 Sulfur distribution and cycling in forest ecosystems. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric
5 deposition and forest nutrient cycling: a synthesis of the integrated forest study. New York, NY:
6 Springer-Verlag; pp. 90-129. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds.
7 Ecological studies: analysis and synthesis: v. 91).
8 Mitchell, M. J.; Foster, N. W.; Shepard, J. P.; Morrison, I. K. (1992b) Nutrient cycling in Huntington Forest and
9 Turkey Lakes deciduous stands: nitrogen and sulfur. Can. J. For. Res. 22: 457-464.
10 Molenar, J. V.; Malm, W. C.; Johnson, C. E. (1994) Visual air quality simulation techniques. Atmos. Environ.
11 28:1055-1063.
12 Monteith, J. L.; Unsworth, M. H. (1990) Principles of environmental physics. 2nd ed. London, United Kingdom:
13 Edward Arnold; pp. 165-175.
14 Moosmtiller, H.; Arnott, W. P.; Rogers, C. F. (1998) Photoacoustic and filter measurements related to aerosol light
15 absorption during the Northern Front Range Air Quality Study (Colorado 1996.1997). J. Geophys. Res.
16 103:28,149-28,157.
17 Moropoulou, A.; Bisbikou, K.; Torts, K.; Van Grieken, R.; Zezza, F.; Macri, F. (1998) Origin and growth of
18 weathering crusts on ancient marbles in industrial atmosphere. Atmos. Environ. 32: 967-982.
19 Mueller, S. F. (1988) Chemical deposition to high-elevation spruce-fir forests in the eastern United States. Muscle
20 Shoals, AL: Tennessee Valley Authority, Air Quality Branch.
21 Naeem, S.; Thompson, L. J.; Lawler, S. P.; Lawton, J. H.; Woodfin, R. M. (1994) Declining biodiversity can alter
22 the performance of ecosystems. Nature 368: 734-737.
23 Nash, T. H., III. (1975) Influence of effluents from a zinc factory on lichens. Ecol. Monogr. 45: 183-198.
24 National Assessment Synthesis Team (NAST). (2000) Climate change impacts on the United States: the potential
25 consequences of climate variability and change. Overview. Washington, DC: U.S. Global Change Research
26 Program.
27 National Park Service. (1998) IMPROVE program goals and objectives. Washington, DC: National Park Service.
28 Available: www2.nature.nps.gov/ard/impr/index.htm [24 November 1999].
29 National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
30 National Academy Press. 3v.
31 National Science and Technology Council. (1998) National acid precipitation assessment program biennial report to
32 Congress: an integrated assessment; executive summary. Silver Spring, MD: U.S. Department of Commerce,
33 National Oceanic and Atmospheric Administration. Available:
34 www.nnic.noaa.gov/CENR/NAPAP/NAPAP_96.htm [24 November 1999].
35 National Weather Service. (1998) Visibility. Washington, DC: National Weather Service, National Oceanic and
36 Atmospheric Administration. Available: www.nws.noaa.gov/asos/vsby.htm [24 November 1999].
37 Nazaroff, W. W.; Cass, G. R. (1991) Protecting museum collections from soiling due to the deposition of airborne
38 particles. Atmos. Environ. Part A 25: 841-852.
39 Nelis, P. M.; Branford, D.; Unsworth, M. H. (1994) A model of the transfer of radioactivity from sea to land in sea
40 spray. Atmos. Environ. 28: 3213-3223.
41 Neuberger, H.; Hosier, C. L.; Kocmond, W. C. (1967) Vegetation as aerosol filter In: Tromp, S. W.; Weihe, W. H.,
42 eds. Biometeorology, v. 2. pp. 693-702.
43 New England Regional Assessment Group (NERAG). (2001) Preparing for a changing climate: the potential
44 consequences of climate variability and change. New England regional overview. Washington, DC: U.S.
45 Environmental Protection Agency; Office of Research and Development; U.S. Global Change Research
46 Program (USGCRP) and Durham, NH: University of New Hampshire; Institute for the Study of Earth,
47 Oceans, and Space.
48 Nicholson, K. W. (1988) The dry deposition of small particles: a review of experimental measurements.
49 Atmos. Environ. 22: 2653-2666.
50 Nicholson, K. W.; Davies, T. D. (1987) Field measurements of the dry deposition of paniculate sulphate.
51 Atmos. Environ. 21: 1561-1571.
52 Nielsen, K. E.; Ladekarl, U. L.; Nornberg, P. (1999) Dynamic soil processes on heathland due to changes in
53 vegetation to oak and Sitka spruce. For. Ecol. Manage. 114: 107-116.
54 Nihlgard, B. (1985) The ammonium hypothesis~an additional explanation to the forest dieback in Europe. Ambio
55 14:2-8.
April 2002 4-230 DRAFT-DO NOT QUOTE OR CITE
-------
1 Nikliriska, M; Laskowski, R.; Maryariski, M. (1998) Effect of heavy metals and storage time on two types of forest
2 litter: basal respiration rate and exchangeable metals. Ecotoxicol. Environ. Safety 41: 8-18.
3 Noble, I. R.; Dirzo, R. (1997) Forests as human-dominated ecosystems. Science (Washington, DC) 277: 522-525.
4 Nodvin, S. C.; Van Miegroet, H.; Lindberg, S. E.; Nicholas, N. S.; Johnson, D. W. (1995) Acidic deposition,
5 ecosystem processes, and nitrogen saturation in a high elevation southern Appalachian watershed. Water Air
6 SoilPollut. 85: 1647-1652.
7 Nord, A. G.; Ericsson, T. (1993) Chemical analysis of thin black layers on building stone. Stud. Conserv. 38: 25-35.
8 Northeast States for Coordinated Air Use Management (NESCAUM). (2002) CAMNET: realtime air pollution and
9 visibility monitoring. Boston, MA: NESCAUM. Available at: http://www.hazecam.net. (23 January 2002).
10 Norton, R. A.; Childers, N. F. (1954) Experiments with urea sprays on the peach. Proc. Am. Soc. Hortic. Sci.
11 63:23-31.
12 Novakov, T.; Penner, J. E. (1993) Large contribution of organic aerosols to cloud-condensation-nuclei
13 concentrations. Nature (London) 365: 823-826.
14 Ockenden, W. A.; Stinnes, E.; Parker, C.; Jones, K. C. (1998) Observations on persistent organic pollutants in
15 plants: implications for their use as passive air samplers and for POP cycling. Environ. Sci. Technol.
16 32:2721-2726.
17 Odum, E. P. (1969) The strategy of ecosystem development: an understanding of ecological succession provides a
18 basis for resolving man's conflict with nature. Science (Washington, DC) 164: 262-270.
19 Odum, E. P. (1985) Trends expected in stressed ecosystems. BioScience 35: 419-422.
20 Odum, E. P. (1993) The ecosystem. In: Ecology and our endangered life-support systems. 2nd ed. Sunderland, MA:
21 Sinauer Associates, Inc.; pp. 38-67.
22 Oehme, M.; Biseth, A.; Schlabach, M.; Wiig, 0. (1995) Concentrations of polychlorinated dibenzo-p-dioxins,
23 dibenzofurans and non-ortho substituted biphenyls in polar bear milk from Svalbard (Norway). Environ.
24 Pollut. 90: 401-407.
25 Olszyk, D. M.; Bytnerowicz, A.; Takemoto, B. K. (1989) Photochemical oxidant pollution and vegetation: effects of
26 gases, fog, and particles. Environ. Pollut. 61: 11-29.
27 Oosting, H. J.; Billings, W. D. (1942) Factors effecting vegetational zonation on coastal dunes. Ecology
28 23: 131-142.
29 Ormrod, D. P. (1984) Impact of trace element pollution on plants. In: Treshow, M., ed. Air pollution and plant life.
30 Chichester, United Kingdom: John Wiley & Sons; pp. 291-319.
31 Ouimette, J. R.; Flagan, R. C. (1982) The extinction coefficient of multicomponent aerosols. Atmos. Environ.
32 16:2405-2420.
33 Padgett, P.; Allen, E. B. (1999) Differential responses to nitrogen fertilization in native shrubs and exotic annuals
34 common to Mediterranean coastal sage scrub of California. Plant Ecol. 144: 93-101.
35 Padgett, P. E.; Allen, E. B.; Bytnerowicz, A.; Minich, R. A. (1999) Changes in soil inorganic nitrogen as related to
36 atmospheric nitrogenous pollutants in southern California. Atmos. Environ. 33: 769-781.
37 Paerl, H. W.; Bales, J. D.; Ausley, L. W.; Buzzelli, C. P.; Crowder, L. B.; Eby, L. A.; Go, M.; Peierls, B. L.;
38 Richardson, T. L.; Ramus, J. S. (2001) Ecosystem impacts of three sequential hurricanes (Dennis, Floyd, and
39 Irene) on the United States' largest lagoonal estuary, Pamlico Sound, NC. Proc. Nat. Acad. Sci. U. S. A.
40 98:5655-5611.
41 Pan, W.; Tatang, M. A.; McRae, G. J.; Prinn, R. G. (1997) Uncertainty analysis of direct radiative forcing by
42 anthropogenic sulfate aerosols. J. Geophys. Res. 102 (D18): 21,915-21,924.
43 Pan, W.; Tatang, M. A.; McRae, G. J.; Prinn, R. G. (1998) Uncertainty analysis of indirect radiative forcing by
44 anthropogenic sulfate aerosols. J. Geophys. Res. 103 (D4): 3815-3823.
45 Pandis, S. N.; Seinfeld, J. H. (1989) Mathematical modeling of acid deposition due to radiation fog. J. Geophys.
46 Res. [Atmos.] 94: 12,911-12,923.
47 Pankow, J. F. (1987) Review and comparative analysis of the theories on partitioning between the gas and aerosol
48 particulate phases in the atmosphere. Atmos. Environ. 21: 2275-2283.
49 Parish, S. B. (1910) The effect of cement dust on citrus trees. Plant World 13: 288-291.
50 Parker, A. (1955) The destructive effects of air pollution on materials. Colliery Guardian 191: 447-450.
51 Parker, G. G. (1983) Throughfall and stemflow in the forest nutrient cycle. In: Macfadyen, A.; Ford, E. D., eds.
52 Advances in ecological research, v. 13; pp. 57-120.
53 Patz, J. A.; Engelberg, D.; Last, J. (2000a) The effects of changing weather on public health. Ann. Rev. Public
54 Health 21: 271-307.
April 2002 4-231 DRAFT-DO NOT QUOTE OR CITE
-------
1 Patz, J. A.; McGeehin, M. A.; Bernard, S. M; Ebi, K. L.; Epstein, P. R.; Grambsch, A.; Gubler, D. I; Reiter, P.;
2 Romieu, L; Rose, J. B.; Samet, J. M.; Trtanj, J. (2000b) The potential health impacts of climate variability
3 and change for the United States: executive summary of the report of the health sector of the U.S. national
4 assessment. Environ. Health Perspect. 108: 367-376.
5 Pearcy, R. W.; Bjorkman, O.; Caldwell, M. M.; Keeley, J. E.; Monson, R. K.; Strain, B. R. (1987) Carbon gain by
6 plants in natural environments. BioScience 37: 21-29.
7 Penndorf, R. B. (1958) An approximation method to the Mie theory for colloidal spheres. J. Phys. Chem. 62:
8 1537-1542.
9 Percy, K. E. (1991) Effects of acid rain on forest vegetation: morphological and non-mensurational growth effects.
10 In: Rennie, P. J.; Robitaille, G., eds. Effects of Acid Rain on Forest Resources: Proceedings of the
11 Conference ; June, 1983; Sainte-Foy, Quebec. Ottawa, Canada: Canadian Forestry Service; pp. 97-111.
12 Perry, K. D.; Cahill, T. A.; Eldred, R. A.; Dutcher, D. D.; Gill, T. E. (1997) Long-range transport of North African
13 dust to the eastern United States. J. Geophys. Res. [Atmos] 102: 11,225-11,238.
14 Peterjohn, W. T.; Adams, M. B.; Gilliam, F. S. (1996) Symptoms of nitrogen saturation in two central Appalachian
15 hardwood forest ecosystems. Biogeochemistry 35: 507-522.
16 Peters, K.; Eiden, R. (1992) Modelling the dry deposition velocity of aerosol particles to a spruce forest.
17 Atmos. Environ. Part A 26: 2555-2564.
18 Peterson, J. T.; Flowers, E. C.; Rudisill, J. H. (1978) Urban-rural solar radiation and atmospheric turbidity
19 measurements in the Los Angeles basin. J. Appl. Meteorol. 17: 1595-1609.
20 Peterson, G.; Allen, C. R.; Holling, C. S. (1998) Ecological resilience, biodiversity, and scale. Ecosystems 1: 6-18.
21 Petuskey, W. T.; Richardson, D. A.; Dolske, D. A. (1995) Aspects of the deterioration of sandstone masonry in
22 Anasazi dwelling ruins at Mesa Verde National Park, Colorado, USA. Sci. Total Environ. 167: 145-159.
23 Piatt, J. R.; Krause, P. D. (1974) Road and site characteristics that influence road salt distribution and damage to
24 roadside aspen trees. Water Air SoilPollut. 3: 301-304.
25 Pickett, S. T. A.; Cadenasso, M. L.; Grove, J. M.; Nilon, C. Y.; Pouyat, R. V.; Zipperer, W. C.; Costanza, R. (2001)
26 Urban ecological systems: linking terrestrial ecological, physical, and socioeconomic components of
27 metropolitan areas. Annu. Rev. Ecol. Syst. 32: 127-157.
28 Pilinis, C.; Pandis, S. N.; Seinfield, J. H. (1995) Sensitivity of direct climate forcing by atmospheric aerosols to
29 aerosol size and composition. J. Geophys. Res. [Atmos.] 100: 18739-18754.
30 Pimentel, D.; Wilson, C.; McCullum, C.; Huang, R.; Dwen, P.; Flack, J.; Tran, Q.; Saltman, T.; Cliff, B. (1997)
31 Economic and environmental benefits of biodiversity. BioScience 47: 747-757.
32 Pimm, S. L. (1984) The complexity and stability of ecosystems. Nature (London) 307: 321-326.
33 Pitchford, M. L.; Malm, W. C. (1994) Development and applications of a standard visual index. Atmos. Environ.
34 28: 1049-1054.
3 5 Pitchford, M. L.; McMurry, P. H. (1994) Relationship between measured water vapor growth and chemistry of
36 atmospheric aerosol for Grand Canyon, Arizona, in winter 1990. Atmos. Environ. 28: 827-839.
37 Pouyat, R. V.; McDonnell, M. J. (1991) Heavy metal accumulations in forest soils along an urban-rural gradient in
3 8 southeastern New York, USA. Water Air Soil Pollut. 57/58: 797-808.
39 Prescott, C. E.; Parkinson, D. (1985) Effects of sulphur pollution on rates of litter decomposition in a pine forest.
40 Can. J. Bot. 63: 1436-1443.
41 Pueschel, R. F. (1993) Potential climatic effects of anthropogenic aerosols. In: Jennings, S. G., ed. Aerosol effects
42 on climate. Tucson, AZ: University of Arizona Press; pp. 110-132.
43 Ragsdale, H. L.; Lindberg, S. E.; Lovett, G. M.; Schaefer, D. A. (1992) Atmospheric deposition and throughfall
44 fluxes of base cations. In: Johnson, D. W.; Lindberg, S. E., eds. Atmospheric deposition and forest nutrient
45 cycling: a synthesis of the integrated forest study. New York, NY: Springer Verlag; pp. 235-253. (Billings,
46 W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis:
47 v. 91).
48 Rapport, D. J.; Whitford, W. G. (1999) How ecosystems respond to stress: common properties of arid and aquatic
49 systems. BioScience 49: 193-203.
50 Raynal, D. J.; Joslin, J. D.; Thornton, F. C.; Schaedle, M.; Henderson, G. S. (1990) Sensitivity of tree seedlings to
51 aluminum: III. Red spruce and loblolly pine. J. Environ. Qual. 19: 180-187.
52 Realini, M.; Negrotti, R.; Appollonia, L.; Vaudan, D. (1995) Deposition of paniculate matter on stone surfaces: an
53 experimental verification of its effects on Carrara marble. In: Saiz-Jimenez, C., ed. The deterioration of
54 monuments: proceedings of the 2nd international symposium on biodeterioration and biodegradation;
55 January 1994; Sevilla, Spain. Sci. Total Environ. 167: 67-72.
April 2002 4-232 DRAFT-DO NOT QUOTE OR CITE
-------
1 Reiners, W. A.; Olson, R. K. (1984) Effects of canopy components on throughfall chemistry: an experimental
2 analysis. Oecologia 63: 320-330.
3 Reisinger, L. M. (1990) Analysis of airborne particles sampled in the southern Appalachian Mountains. Water Air
4 SoilPollut. 50: 149-162.
5 Rennie, P. J. (1955) The uptake of nutrients by mature forest growth. Plant Soil 7: 49-95.
6 Reuss, J. O. (1983) Implications of the calcium-aluminum exchange system for the effect of acid precipitation on
7 soils. J. Environ. Qual. 12: 591-595.
8 Reuss, J. O.; Johnson, D. W. (1986) Acid deposition and the acidification of soils and waters. New York, NY:
9 Springer-Verlag. (Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. S.; Remmert, H., eds. Ecological
10 studies: analysis and synthesis: v. 59).
11 Riggan, P. J.; Lockwood, R. N.; Lopez, E. N. (1985) Deposition and processing of airborne nitrogen pollutants in
12 Mediterranean-type ecosystems of southern California. Environ. Sci. Technol. 19: 781-789.
13 Rodney, D. R. (1952) The entrance of nitrogen compounds through the epidermis of apple leaves. Proc. Am. Soc.
14 Hortic. Sci. 59: 99-102.
15 Roelofs, J. G. M.; Kempers, A. J.; Houdijk, A. L. F. M.; Jansen, J. (1985) The effect of air-borne ammonium
16 sulphate on Pinus nigra var. maritima in the Netherlands. Plant Soil 84: 45-56.
17 Roelofs, J. G. M.; Boxman, A. W.; VanDijk, H. F. G. (1987) Effects of airborne ammonium on natural vegetation
18 and forests. In: Asman, W. A. H.; Diederen, H. S. M. A., eds. Ammonia and acidification: proceedings [of a]
19 symposium of the European Association for the Science of Air Pollution (EURAS AP); April; Bilthoven,
20 The Netherlands. European Association for the Science of Air Pollution; pp. 266-276.
21 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993a) Sources of fine organic
22 aerosol. 4. Paniculate abrasion products from leaf surfaces of urban plants. Environ. Sci. Technol.
23 27:2700-2711.
24 Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.; Simoneit, B. R. T. (1993b) Sources of fine organic
25 aerosol. 3. Road dust, tire debris, and organometallic brake lining dust: roads as sources and sinks.
26 Environ. Sci. Technol. 27: 1892-1904.
27 Rosenberg, N. J.; Blad, B. L.; Verma, S. B. (1983) Microclimate: the biological environment. 2nd ed. New York,
28 NY: John Wiley & Sons.
29 Rovira, A. D.; Davey, C. B. (1974) Biology of the rhizosphere. In: Carson, E. W., ed. The plant root and its
30 environment: proceedings of an institute; July, 1971; Blacksburg, VA. Charlottesville, VA: University Press
31 of Virginia; pp. 153-204.
3 2 Rueth, H.; Baron, J. S. (2002) Differences in Englemann spruce forest biogeochemistry east and west of the
33 Continental Divide in Colorado, USA. Ecosystems 5: 45-57.
34 Sabbioni, C. (1995) Contribution of atmospheric deposition to the formation of damage layers. In: Jimenez-Saiz, C.,
3 5 ed. The deterioration of monuments: proceedings of the 2nd international symposium on biodeterioration and
36 biodegradation; January 1994; Sevilla, Spain. Sci. Total Environ. 167: 49-56.
37 Sabbioni, C.; Zappia, G.; Gobbi, G. (1992) Carbonaceous particles on carbonate building stones in a simulated
38 system. In: Proceedings of the 1992 European aerosol conference; September; Oxford, United Kingdom.
39 J. Aerosol Sci. 23(suppl. 1): S921-S924.
40 Sabbioni, C.; Zappia, G.; Gobbi, G. (1996) Carbonaceous particles and stone damage in a laboratory exposure
41 system. J. Geophys. Res. [Atmos.] 101: 19,621-19,627.
42 Sabbioni, C.; Zappia, G.; Ghedini, N.; Gobbi, G.; Favoni, O. (1998) Black crusts on ancient mortars.
43 Atmos. Environ. 32: 215-223.
44 Saiz-Jimenez, C. (1993) Deposition of airborne organic pollutants on historic buildings. Atmos. Environ. Part B
45 27:77-85.
46 Samecka-Cymerman, A.; Kempers, A. J. (1999) Bioindication of heavy metals in the town Wro/caw (Poland) with
47 evergreen plants. Atmos. Environ. 33: 419-430.
48 Sanyal, B.; Singhania, G. K. (1956) Atmospheric corrosion of metals: part I. J. Sci. Ind. Res. Sect. B 15: 448-455.
49 Saunders, P. J. W.; Godzik, S. (1986) Terrestrial vegetation-air pollutant interactions: nongaseous air pollutants.
50 In: Legge, A. H.; Krupa, S. V., eds. Air pollutants and their effects on the terrestrial ecosystem. New York,
51 NY: John Wiley & Sons; pp. 389-394. (Advances in environmental science and technology: v. 18).
52 Saxena, P.; Peterson, T. W. (1981) Thermodynamics of multicomponent electrolytic aerosols. J. Colloid Interface
53 Sci. 79:496-510.
54 Saxena, P.; Mueller, P. K.; Kim, Y. P.; Seinfeld, J. H.; Koutrakis, P. (1993) Coupling thermodynamic theory with
55 measurements to characterize acidity of atmospheric particles. Aerosol Sci. Technol. 19: 279-293.
April 2002 4-233 DRAFT-DO NOT QUOTE OR CITE
-------
1 Saxena, P.; Hildemann, L. M; McMurry, P. H.; Seinfeld, J. H. (1995) Organics alter hygroscopic behavior of
2 atmospheric particles. J. Geophys. Res. [Atmos.] 100: 18,755-18,770.
3 Schiavon, N.; Chiavari, G.; Schiavon, G.; Fabbri, D. (1995) Nature and decay effects of urban soiling on granitic
4 building stones. In: Jimenez-Saiz, C., ed. The deterioration of monuments: proceedings of the 2nd
5 international symposium on biodeterioration and biodegradation; January 1994; Sevilla, Spain. Sci. Total
6 Environ. 167: 87-101.
7 Schichtel, B. A.; Husar, R. B.; Falke, S. R.; Wilson, W. E. (2001) Haze trends over the United States, 1980-1995.
8 Atmos. Environ. 35: 5205-5210.
9 Schier, G. A.; Jensen, K. (1992) Atmospheric deposition effects on foliar injury and foliar leaching in red spruce.
10 In: Eager, C.; Adams, M. B., eds. Ecology and decline of red spruce in the eastern United States. New York,
11 NY: Springer-Verlag; pp. 271-294. [Billings, W. D.; Golley, F.; Lange, O. L.; Olson, J. D.; Remmert, H.,
12 eds. Ecological Studies, Analysis and Synthesis Ecological Studies, v. 96].
13 Schimek, M. G. (1981) Smoothing and regression. New York, NY: Wiley.
14 Schnug, E. (1997) Significance of sulphur for the quality of domesticated plants. In: Cram, W. J.; De Kok, L. J.;
15 Stolen, L; Brunold, C.; Rennenberg, H., eds. Sulphur metabolism in higher plants: molecular,
16 ecophysiological and nutritional aspects. Leiden, The Netherlands: Backhuys Publishers; pp. 109-130.
17 Schonherr, J.; Huber, R. (1977) Plant cuticles are polyelectrolytes with isoelectric points around three. Plant
18 Physiol. 59: 145-150.
19 Schulze, E.-D. (1989) Air pollution and forest decline in a spruce (Picea abies) forest. Science (Washington, DC)
20 244:776-783.
21 Schuster, P. F.; Reddy, M. M.; Sherwood, S. I. (1994) Effects of acid rain and sulfur dioxide on marble dissolution.
22 Mater. Perform. 33:76-80.
23 Schwar, M. J. R. (1998) Nuisance dust deposition and soiling rate measurements. Environ. Technol. 19: 223-229.
24 Seckmeyer, G.; McKenzie, R. L. (1992) Increased ultraviolet radiation in New Zealand (45 degrees S) relative to
25 Germany (48 degrees N). Nature 359: 135-137.
26 Sehmel, G. A. (1980) Particle and gas dry deposition: a review. Atmos. Environ. 14: 983-1011.
27 Sehmel, G. A.; Hodgson, W. H. (1976) Predicted dry deposition velocities. In: Atmosphere-surface exchange of
28 paniculate and gaseous pollutants (1974): proceedings of a symposium; September 1974; Richland, WA.
29 Oak Ridge, TN: Energy Research and Development Administration; pp. 399-422. (ERDA symposium series:
30 38). Available from: NTIS, Springfield, VA; CONF-740921.
31 Seinfeld, J. H. (1989) Urban air pollution: state of the science. Science 243: 745-752.
32 Sereda, P. J. (1974) Weather factors affecting corrosion of metals. In: Corrosion in natural environments: three
3 3 symposia presented at the seventy-sixth annual meeting [of the] American Society for Testing and Materials;
34 June 1973; Philadelphia, PA. Philadelphia, PA: American Society for Testing and Materials; pp. 7-22.
3 5 (ASTM special technical publication 558).
36 Shannon, J. D. (1999) Regional trends in wet deposition of sulfate in the United States and SO2 emissions from
37 1980 through 1995. Atmos. Environ. 33: 807-816.
3 8 Shinn, J. H. (1978) A critical survey of measurements of foliar deposition of airborne sulfates and nitrates.
39 Presented at: 71st annual meeting of the Air Pollution Control Association; June; Houston, TX. Pittsburgh,
40 PA: Air Pollution Control Association; paper no. 78-7.2.
41 Shortle, W. C.; Bondietti, E. A. (1992) Timing, magnitude, and impact of acidic deposition on sensitive forest sites.
42 Water Air Soil Pollut. 61: 253-267.
43 Shortle, W. C.; Smith, K. T. (1988) Aluminum-induced calcium deficiency syndrome in declining red spruce.
44 Science (Washington, DC) 240: 1017-1018.
45 Shortle, W. C.; Smith, K. T.; Minocha, R.; Lawrence, G. B.; David, M. B. (1997) Acidic deposition, cation
46 mobilization, and biochemical indicators of stress in healthy red spruce. J. Environ. Qual. 26: 871-876.
47 Showak, W.; Dunbar, S. R. (1982) Effect of 1 percent copper addition on atmospheric corrosion of rolled zinc after
48 20 years' exposure. In: Dean, S. W.; Rhea, E. C., eds. Atmospheric corrosion of metals: a symposium;
49 May 1980; Denver, CO. Philadelphia, PA: American Society for Testing and Materials; pp. 135-162.
50 (ASTM special technical publication 767).
51 Shriner, D. S.; Henderson, G. S. (1978) Sulfur distribution and cycling in a deciduous forest watershed. J. Environ.
52 Qual. 7: 392-397.
53 Sievering, H. (1987) Dynamics of sulfur exchange at the air/forest canopy interface: a review of throughfall inferred
54 deposition rates. Global Biogeochem. Cycles 1: 233-249.
April 2002 4-234 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sievering, H.; Rusch, D.; Marquez, L. (1996) Nitric acid, paniculate nitrate and ammonium in the continental free
2 troposphere: nitrogen deposition to an alpine tundra ecosystem. Atmos. Environ. 30: 2527-2537.
3 Simonich; S. L.; Hites, R. A. (1995) Organic pollutant accumulation in vegetation. Environ. Sci. Technol.
4 29:2905-2914.
5 Simpson, J. W.; Horrobin, P. J., eds. (1970) The weathering and performance of building materials. New York, NY:
6 Wiley-Interscience.
7 Sisler, J. F. (1996) Spatial and seasonal patterns and long term variability of the composition of the haze in the
8 United States: an analysis of data from the IMPROVE network. Fort Collins, CO: Colorado State University,
9 Cooperative Institute for Research in the Atmosphere; EPA air docket A096-56, document no. VI-B-09-(ee).
10 Sisler, J. F.; Cahill, T. A. (1993) Spacial and temporal patterns and the chemical composition of the haze and its
11 impact on visibility in Alaska. In: Proceedings [of the] 86th annual meeting & exhibition of the Air & Waste
12 Management Association. Volume 8: general environmental topics; June; Denver, CO. Pittsburgh, PA: Air &
13 Waste Management Association; 93-MP-4.03.
14 Sisler, J. F.; Malm, W. C. (2000) Interpretation of trends of PM25 and reconstructed visibility from the IMPROVE
15 network. J. Air Waste Manage. Assoc. 50: 775-789.
16 Sisler, J. F.; Huffman, D.; Lattimer, D. A.; Malm, W. C.; Pitchford, M. L. (1993) Spatial and temporal patterns and
17 the chemical composition of the haze in the United States: an anlysis of data from the IMPROVE network,
18 1988-1991. Fort Collins, CO: Colorado State University, Cooperative Institute for Research in the
19 Atmosphere (CIRA).
20 Skeffington, R. A.; Wilson, E. J. (1988) Excess nitrogen deposition: issues for consideration. Environ. Pollut.
21 54: 159-184.
22 Skerry, B. S.; Alavi, A.; Lindgren, K. I. (1988) Environmental and electrochemical test methods for the evaluation
23 of protective organic coatings. J. Coat. Technol. 60: 97-106.
24 Slinn, W. G. N. (1977) Some approximations for the wet and dry removal of particles and gases from the
25 atmosphere. Water Air Soil Pollut. 7: 513-543.
26 Slinn, W. G. N. (1978) Parameterizations for resuspension and for wet and dry deposition of particles and gases for
27 use in radiation dose calculations. Nucl. Saf. 19: 205-219.
28 Slinn, W. G. N. (1982) Predictions for particle deposition to vegetative canopies. Atmos. Environ. 16: 1785-1794.
29 Sloane, C. S. (1983) Optical properties of aerosols-comparison of measurements with model calculations.
30 Atmos. Environ. 17: 409-416.
31 Sloane, C. S. (1984) Optical properties of aerosols of mixed composition. Atmos. Environ. 18: 871-878.
32 Sloane, C. S. (1986) Effect of composition on aerosol light scattering efficiencies. Atmos. Environ. 20: 1025-1037.
33 Sloane, C. S.; Wolff, G. T. (1985) Prediction of ambient light scattering using a physical model responsive to
34 relative humidity: validation with measurements from Detroit. Atmos. Environ. 19: 669-680.
35 Smith, W. H. (1973) Metal contamination of urban woody plants. Environ. Sci. Technol. 7: 631-636.
3 6 Smith, W. H. (1974) Air pollution - effects on the structure and function of the temperate forest ecosystem.
37 Environ. Pollut. 6: 111-129.
3 8 Smith, W. H. (1984) Pollutant uptake by plants. In: Treshow, M., ed. Air pollution and plant life. New York, NY:
39 John Wiley & Sons, Ltd.; pp. 417-450.
40 Smith, W. H. (1990a) Forests as sinks for air contaminants: soil compartment. In: Air pollution and forests:
41 interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag;
42 pp. 113-146. (Springer series on environmental management).
43 Smith, W. H. (1990b) Forests as sinks for air contaminants: vegetative compartment. In: Air pollution and forests:
44 interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag;
45 pp. 147-180. (Springer series on environmental management).
46 Smith, W. H. (1990c) Forest nutrient cycling: toxic ions. In: Air pollution and forests: interactions between air
47 contaminants and forest ecosystems. 2nd ed. New York, NY: Springer-Verlag; pp. 225-268. (Springer series
48 on environmental management).
49 Smith, W. H. (1990d) Forest biotic agent stress: air pollutants and disease caused by microbial pathogens. In: Air
50 pollution and forests: interactions between air contaminants and forest ecosystems. 2nd ed. New York, NY:
51 Springer-Verlag; pp. 366-397. (Springer series on environmental management).
52 Smith, W. H. (1991) Air pollution and forest damage. Chem. Eng. News 69(45): 30-43.
53 Smith, K. E. C.; Jones, K. C. (2000) Particles and vegetation: implications for the transfer of particle-bound organic
54 contaminants to vegetation. Sci. Total Environ. 246: 207-236.
April 2002 4-23 5 DRAFT-DO NOT QUOTE OR CITE
-------
1 Smith, W. H.; Staskawicz, B. J. (1977) Removal of atmospheric particles by leaves and twigs of urban trees:
2 some preliminary observations and assessment of research needs. Environ. Manage. (N. Y.) 1: 317-330.
3 Soulen, P. E.; Frederick, J. E. (1999) Estimating biologically active UV irradiance from satellite radiance
4 measurements: a sensitivity study. J. Geophys. Res. 104: 4117-4126.
5 Spence, J. W.; Haynie, F. H. (1972) Paint technology and air pollution: a survey and economic assessment.
6 Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Programs; publication no.
7 AP-103. Available from: NTIS, Springfield, VA; PB-210736.
8 Spence, J. W.; Haynie, F.; Upham, J. B. (1975) Effects of gaseous pollutants on paints: a chamber study. J. Paint
9 Technol. 47: 57-63.
10 Spence, J. W.; Haynie, F. H.; Lipfert, F. W.; Cramer, S. D.; McDonald, L. G. (1992) Atmospheric corrosion model
11 for galvanized steel structures. Corrosion 48: 1009-1019.
12 Spiker, E. C.; Comer, V. J.; Hosker, R. P.; Sherwood, S. I. (1992) Dry deposition of SO2 on limestone and marble:
13 the role of humidity. In: Rodrigues, J. D.; Henriques, F.; Jeremias, F. T., eds. Proceedings of the 7th
14 international congress on deterioration and conservation of stone; June; Lisbon, Portugal. Lisbon, Portugal:
15 Laboratorio Nacional de Engenharia Civil; pp. 397-406.
16 Spinka, J. (1971) Vliv znecisteneho ovzdusi na ovocne stromy azeleninu [Effects of polluted air on fruit trees and
17 legumes]. Ziva 19: 13-15.
18 Steubing, L.; Klee, R. (1970) Vergleichende Untersuchungen zur Staubfilterwirkung von Laub- und Nadelgeholzen
19 [Comparative investigations on the dust-filter effect of broad-leaved and coniferous trees]. Angew. Bot.
20 44:73-85.
21 Stevens, R. K.; Dzubay, T. G.; Lewis, C. W.; Shaw, R. W., Jr. (1984) Source apportionment methods applied to the
22 determination of the origin of ambient aerosols that affect visibility in forested areas. Atmos. Environ.
23 18:261-272.
24 Stoddard, J. L. (1994) Long-term changes in watershed retention of nitrogen: its causes and aquatic consequences.
25 In: Baker, L. A., ed. Environmental chemistry of lakes and reservoirs. Washington, DC: American Chemical
26 Society; pp. 223-284. (Advances in chemistry series no. 237).
27 Stoddard, J. L.; Murdoch, P. S. (1991) Catskill Mountains. In: Charles, D. F., ed. Acidic deposition and aquatic
28 ecosystems: regional case studies. New York, NY: Springer-Verlag; pp. 237-271.
29 Storm, G. L.; Fosmire, G. J.; Bellis, E. D. (1994) Persistence of metals in soil and selected vertebrates in the vicinity
30 of the Palmerton zinc smelters. J. Environ. Qual. 23: 508-514.
31 Strandberg, H.; Johansson, L.-G. (1997) The formation of black patina on copper in humid air containing traces of
32 SO2. J. Electrochem. Soc. 144: 81-89.
33 Suzuki, D. T. (1997) Barriers to perception: from a world of interconnection to fragmentation. In: Raven, P. H., ed.
34 Nature and Human Society: the quest for a sustainable world. Washington, DC: National Academy Press;
35 pp. 11-21.
36 Swietlicki, E.; Zhou, J. C.; Berg, O. H.; Martinsson, B. G.; Frank, G.; Cederfelt, S. L; Desek, U.; Berner, A.;
37 Birmilli, W.; Wiedensohler, A.; Yuskiewicz, B.; Bower, K. N. (1999) A closure study of sub-micrometer
38 aerosol particle hygroscopic behaviour. Atmos. Res. 50: 205-240.
39 Swietlik, D.; Faust, M. (1984) Foliar nutrition of fruit crops. Hortic. Rev. 6: 287-355.
40 Sydberger, T.; Ericsson, R. (1977) Laboratory testing of the atmospheric corrosion of steel. Werkst. Korros.
41 28: 154-158.
42 Taback, H. J.; Brienza, A. R.; Macko, J.; Brunetz, N. (1979) Fine particle emissions from stationary and
43 miscellaneous sources in the south coast air basin: appendix. Sacramento, CA: California Air Resources
44 Board; contract no. A6-191-30.
45 Talbot, J. J. (1979) A review of the potential biological impacts of cooling tower salt drift. Atmos. Environ.
46 13:395-405.
47 Tang, I. N. (1980) Deliquescence properties and particle size change of hygroscopic aerosols. In: Willeke, K., ed.
48 Generation of aerosols and facilities for exposure experiments. Ann Arbor, MI: Ann Arbor Science
49 Publishers, Inc.; pp. 153-167.
50 Tang, I. N. (1997) Thermodynamic and optical properties of mixed-salt aerosols of atmospheric importance.
51 J. Geophys. Res. [Atmos.] 102: 1883-1893.
52 Tang, I. N.; Munkelwitz, H. R. (1994) Water activities, densities, and refractive indices of aqueous sulfates and
53 sodium nitrate droplets of atmospheric importance. J. Geophys. Res. [Atmos.] 99: 18,801-18,808.
54
55
April 2002 4-236 DRAFT-DO NOT QUOTE OR CITE
-------
1 Tanner, R. L. (1981) An ambient experimental study of phase equilibrium in the atmospheric system: aerosol H+,
2 NH4+, SO42", NO3" ~NH3(g), HNO3(g). In: Fogler, H. S., ed. Chemical reactors: based on a symposium
3 sponsored by the Division of Industrial and Engineering Chemistry at the second chemical congress of the
4 North American continent (180th ACS national meeting); August 1980; Las Vegas, NV. Washington, DC:
5 American Chemical Society; pp. 527-532. (ACS symposium series 168).
6 Taylor, G. E., Jr.; Hanson, P. J.; Baldocchi, D. D. (1988) Pollutant deposition to individual leaves and plant
7 canopies: sites of regulation and relationship to injury. In: Heck, W. W.; Taylor, O. C.; Tingey, D. T., eds.
8 Assessment of crop loss from air pollutants. New York, NY: Elsevier Applied Science; pp. 227-257.
9 Tegen, L; Lacis, A. A. (1996) Modelling of particle size distribution and its influence on the radiative properties of
10 mineral dust aerosol. J. Geophys. Res. Atmos. 101: 19,237-19,244.
11 Tegen, L; Lacis, A. A.; Fung, I. (1996) The influence on climate forcing of mineral aerosols from disturbed soils.
12 Nature (London) 380: 419-422.
13 Thornton, F. C.; Schaedle, M; Raynal, D. J. (1987) Effects of aluminum on red spruce seedlings in solution culture.
14 Environ. Exp. Bot. 27: 489-498.
15 Tilman, D. (1996) Biodiversity: population versus ecosystem stability. Ecology 77: 350-363.
16 Tilman, D. (2000) Causes, consequences and ethics of biodiversity. Nature 405: 208-211.
17 Tilman, D.; Downing, J. A. (1994) Biodiversity and stability in grasslands. Nature (London) 367: 363-365.
18 Tjepkema, J. D.; Cartica, R. J.; Hemond, H. F. (1981) Atmospheric concentration of ammonia in Massachusetts and
19 deposition on vegetation. Nature (London) 294: 445-446.
20 Tombach, I.; Thurston, S. A. (1994) The quality of the SCENES measurements: the roles of data quality goals and
21 evolving technology. In: Aerosols and atmospheric optics: radiative balance and visual air quality:
22 proceedings of the international specialty conference, volume A; September; Snowbird, UT. Pittsburgh, PA:
23 Air & Waste Management Association; pp. 21-32. (A&WMA publication VIP-41).
24 Tong, S. T. Y. (1991) The retention of copper and lead paniculate matter in plant foliage and forest soil.
25 Environ. Int. 17:31-37.
26 Trettin, C.C.; Johnson, D.W.; Todd, D.E., Jr. (1999) Forest nutrient and carbon pools at Walker Branch watershed:
27 changes during a 21-year-period. Soil Sci. Soc. Amer. J. 63: 1436-1448.
28 Trijonis, J.; Shapland, D. (1979) Existing visibility levels in the United States: isopleth maps of visibility in
29 suburban/nonurban areas during 1974-76. Research Triangle Park, NC: U.S. Environmental Protection
30 Agency, Office of Air Quality Planning and Standards; report no. EPA-450/5-79-010. Available from:
31 NTIS, Springfield, VA; PB80-156177.
32 Trijonis, J. C.; Pitchford, M.; McGown, M. (1987) Preliminary extinction budget results from the RESOLVE
33 program. In: Bhardwaja, P. S., ed. Visibility protection: research and policy aspects, an APCA international
34 specialty conference; September 1986; Grand Teton National Park, WY. Pittsburgh, PA: Air Pollution
35 Control Association; pp. 872-883. (APCA transactions series no. TR-10).
36 Trijonis, J. C.; Malm, W. C.; Pitchford, M.; White, W. H.; Charlson, R.; Husar, R. (1991) Visibility: existing and
37 historical conditions-causes and effects. In: Irving, P. M., ed. Acidic deposition: state of science and
38 technology, volume III: terrestrial, materials, health and visibility effects. Washington, DC: The U.S.
39 National Acid Precipitation Assessment Program. (State of science and technology report no. 24).
40 Turner, J.; Kelly, J. (1981) Relationships between soil nutrients and vegetation in a north coast forest, New South
41 Wales. Aust. For. Res. 11: 201-208.
42 Turner, J.; Lambert, M. J. (1980) Sulfur nutrition of forests. In: Shriner, D. S.; Richmond, C. R.; Lingberg, S. E.,
43 eds. Atmospheric sulfur deposition: environmental impact and health effects, proceedings of the second life
44 sciences symposium; October 1979; Gatlinburg, TN. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.;
45 pp. 321-333.
46 Turner, J.; Lambert, M. J.; Gessel, S. P. (1977) Use of foliage sulphate concentrations to predict response to urea
47 application by Douglas-fir. Can. J. For. Res. 7: 476-480.
48 Twomey, S. (1974) Pollution and the planetary albedo. Atmos. Environ. 8: 1251-1256.
49 Twomey, S. (1977) Atmospheric aerosols. Amsterdam, The Netherlands: Elsevier Scientific Publishing Company;
50 p. 237.
51 Tyler, G. (1972) Heavy metals pollute nature, may reduce productivity. Ambio 1: 52-59.
52 U.S. Code. (1977) Clean Air Act, "section" 169A, visibility protection for federal class I areas. U. S. C.
53 42: "section"7491.
54 U.S. Code. (1990) Clean Air Act, "section" 169B, visibility. U. S. C. 42: "section"7492.
April 2002 4-237 DRAFT-DO NOT QUOTE OR CITE
-------
1 U.S. Code. (1991) Clean Air Act, section 108, air quality criteria and control techniques, section 109, national
2 ambient air quality standards. U. S. C. 42: sections 7408-7409.
3 U.S. Environmental Protection Agency. (1979) Protecting visibility: an EPA report to Congress. Research Triangle
4 Park, NC: Office of Air Quality Planning and Standards; EPA report no. EPA-450/5-79-008. Available from:
5 NTIS, Springfield, VA; PB80-220320.
6 U.S. Environmental Protection Agency. (1982) Air quality criteria for paniculate matter and sulfur oxides. Research
7 Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
8 Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
9 U.S. Environmental Protection Agency. (1986 ) Air quality criteria for lead. Research Triangle Park, NC: Office of
10 Health and Environmental Assessment, Environmental Criteria and Assessment Office; EPA report no.
11 EPA-600/8-83/028aF-dF. 4v. Available from: NTIS, Springfield, VA; PB87-142378.
12 U.S. Environmental Protection Agency. (1987) Assessing the risks of trace gases that can modify the stratosphere.
13 Washington, DC: U.S. Environmental Protection Agency, Office of Air Radiation; report no. EPA
14 400/1-87/001A-H.
15 U.S. Environmental Protection Agency. (1993) Air quality criteria for oxides of nitrogen. Research Triangle Park,
16 NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
17 nos. EPA/600/8-9!/049aF-cF. 3v. Available from: NTIS, Springfield, VA; PB95-124533, PB95-124525, and
18 PB95-124517.
19 U.S. Environmental Protection Agency. (1995a) Interim findings on the status of visibility research. Research
20 Triangle Park, NC: Office of Research and Development; report no. EPA/600/R-95/021.
21 U.S. Environmental Protection Agency. (1995b) Testing of meteorological and dispersion models foruse in
22 regional air quality modeling. Research Triangle Park, NC: Office of Air Quality Planning and Standards;
23 report no. EPA-454/R-95-005. Available from: NTIS, Springfield, VA; PB95-215638.
24 U.S. Environmental Protection Agency. (1996a) Air quality criteria for paniculate matter. Research Triangle Park,
25 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
26 U.S. Environmental Protection Agency. (1996b) Air quality criteria for ozone and related photochemical oxidants,
27 v. I-III. Research Triangle Park, NC: Office of Research and Development; report no.
28 EPA/600/P-93/004aF-cF. Available from: NTIS, Springfield, VA; PB96-185582. PB96-185590, and
29 PB96-185608. Available: www.epa.gov/ncea/ozone.htm [25 February 2002].
30 U.S. Environmental Protection Agency. (1997a) Nitrogen oxides: impacts on public health and the environment.
31 Washington, DC: Office of Air and Radiation; August. Available:
32 www.epa.gov/ttncaaal/tl/reports/noxrept.pdf [1999, November 24].
33 U.S. Environmental Protection Agency. (1997b) Summary of the paniculate monitoring program. Washington, DC:
3 4 Office of Air and Radiation.
35 U.S. Environmental Protection Agency. (1998) Monitoring PM2.5 in ambient air using designated reference of class
36 I equivalent methods. In: Quality Assurance Handbook for Air Pollution Measurement Systems, V. II, Part
37 II, Section 2.12. Research Triangle Park, NC: Research and Development, Environmental Monitoring
3 8 Systems Laboratory.
39 U.S. Environmental Protection Agency. (1999a) Protection of visibility. C. F. R. 40: "Section" 51.300-51.309.
40 U.S. Environmental Protection Agency. (1999b) Visibility monitoring guidance. Research Triangle Park, NC:
41 Emissions, Monitoring, and Analysis Division.
42 U.S. Environmental Protection Agency. (2000a) Deposition of air pollutants to the great waters. Third report to
43 Congress. [Executive Summary]. Research Triangle Park, NC: U.S. Environmental Protection Agency,
44 Office of Air Quality Planning and Standards; report no. EPA-453/R-00-005.
45 U.S. Environmental Protection Agency. (2000b) Latest findings on national air quality: 1999 status and trends.
46 Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and
47 Standards; report no. EPA-454/F-00-002.
48 U.S. Environmental Protection Agency. (2001) National air quality and emissions trends report, 1999. Research
49 Triangle Park, NC: Office of Air Quality Planning and Standards; report no. EPA/454/R-01-004. Available:
50 http://www.epa.gov/oar/aqtrnd99/ [11 March, 2002].
51 Ulrich, B. (1983) Interaction of forest canopies with atmospheric constituents: SO2, alkali and earth alkali cations
52 and chloride. In: Ulrich, B.; Pankrath, J., eds. Effects of accumulation of air pollutants in forest ecosystems:
53 proceedings of a workshop; May 1982; Gottingen, Federal Republic of Germany. Dordrecht, The
54 Netherlands: D. Reidel Publishing Company; pp. 33-45.
April 2002 4-23 8 DRAFT-DO NOT QUOTE OR CITE
-------
1 United Nations Environment Programme (UNEP). (1986) Report of the international conference on the assessment
2 of the role of carbon dioxide and of other greenhouse gases in climate variations and associated impacts;
3 October 1985; Villach, Austria. Geneva, Switzerland: World Meteorological Organization; WMO no. 661.
4 United Nations Environment Programme (UNEP). (1998) Environmental effects of ozone depletion: 1998
5 assessment. J. Photochem. Photobiol. B 46: 1-4.
6 United Nations Environment Programme (UNEP). (2000) Environmental effects of ozone depletion: interim
7 summary. Available at: http://www.gcrio.org/ozone/unep2000summarv.html [9 April 20021.
8 Unsworth, M. H. (1984) Evaporation from forests in cloud enhances the effects of acid deposition. Nature (London)
9 312:262-264.
10 Unsworth, M. H.; Wilshaw, J. C. (1989) Wet, occult and dry deposition of pollutants onforests. Agric. For.
11 Meteorol. 47: 221-238.
12 Van Aalst, R. M. (1982) Dry deposition of NOx. In: Schneider, T.; Grant, L., eds. Air pollution by nitrogen oxides.
13 Amsterdam, The Netherlands: Elsevier Scientific Publishing Company; pp. 263-270.
14 Van Breemen, N.; Van Dijk, H. F. G. (1988) Ecosystem effects of atmospheric deposition of nitrogen in The
15 Netherlands. In: Dempster, J. P.; Manning, W. J., eds. Excess nitrogen deposition. Environ. Pollut.
16 54:249-274.
17 VanBreemen, N.; Burrough, P. A.; Velthorst, E. J.; VanDobben, H. F.; De Wit, T.; Ridder, T. B.; Reijnders, H. F.
18 R. (1982) Soil acidification from atmospheric ammonium sulphate in forest canopy throughfall. Nature
19 (London) 299: 548-550.
20 Van Breemen, N.; Mulder, J.; Driscoll, C. T. (1983) Acidification and alkalinization of soils. Plant Soil
21 75:283-308.
22 Van de Hulst, H. C. (1957, reprint 1981) Light scattering by small particles. New York, NY: Wiley. (Structure of
23 matter series).
24 Van der Leun, J. C.; Tang, X.; Tevini, M. (1995) Environmental effects of ozone depletion: 1994 assessment.
25 Ambio 24: 138.
26 Van der Leun, J. C.; Tang, X.; Tevini, M., eds. (1998) Environmental effects of ozone depletion: 1998 assessment.
27 J. Photochem. Photobiol. B: Biology 46(1-3): 1-108.
28 Van Dijk, H. F. G.; Roelofs, J. G. M. (1988) Effects of excessive ammonium deposition on the nutritional status
29 and condition of pine needles. Physiol. Plant. 73: 494-501.
30 Van Dorland, R.; Dentener, F. J.; Lelieveld, J. (1997) Radiative forcing due to tropospheric ozone and sulfate
31 aerosols. J. Geophys. Res. [Atmos.] 102: 28,079-28,100.
32 Van Miegroet, H.; Cole, D. W. (1984) The impact of nitrification on soil acidification and cation leaching in red
33 alder ecosystem. J. Environ. Qual. 13: 586-590.
34 Van Miegroet, H.; Johnson, D. W.; Todd, D. E. (1993) Foliar response of red spruce saplings to fertilization with
35 Ca and Mg in the Great Smoky Mountains National Park. Can. J. For. Res. 23: 89-95.
36 Venkatram, A.; Pleim, J. (1999) The electrical analogy does not apply to modeling dry deposition of particles.
37 Atmos. Environ. 33: 3075-3076.
38 Viles, H. A. (1990) The early stages of building stone decay in an urban environment. Atmos. Environ.
39 24A: 229-232.
40 Viskari, E.-L.; Karenlampi, L. (2000) Roadside Scots pine as an indicator of deicing salt use - a comparative study
41 from two consecutive winters. Water Air Soil Pollut. 122: 405-419.
42 Vitousek, P. M.; Ehrlich, P. R.; Chrlich, A. H.; Matson, P. A. (1986) Human appropriation of the products of
43 photosynthesis. Bioscience 36: 368-373.
44 Vitousek, P. M.; Mooney, H. A.; Lubchenco, J.; Melillo, J. M. (1997) Human domination of Earth's ecosystems.
45 Science (Washington, DC) 277: 494-499.
46 Vogt, K. A.; Dahlgren, R.; Ugolini, F.; Zabowski, D.; Moore, E. E.; Zasoski, R. (1987a) Aluminum, Fe, Ca, Mg, K,
47 Mn, Cu, Zn and P in above- and belowground biomass. I. Abies amabilis and Tsuga mertensiana.
48 Biogeochemistry 4: 277-294.
49 Vogt, K. A.; Dahlgren, R.; Ugolini, F.; Zabowski, D.; Moore, E. E.; Zasoski, R. (1987b) Aluminum, Fe, Ca, Mg, K,
50 Mn, Cu, Zn and P in above- and belowground biomass. II. Pools, and circulation in a subalpine Abies
51 amabilis stand. Biogeochemistry 4: 295-311.
52 Von Glasow, R.; Bott, A. (1999) Interaction of radiation fog with tall vegetation. Atmos. Environ. 33: 1333-1346.
53 Vora, A. B.; Bhatnagar, A. R. (1987) Comparative study of dust fall on the leaves in high pollution and low
54 pollution area of Ahmedabad: V. Caused foliar injury. J. Environ. Biol. 8: 339-346.
April 2002 4-239 DRAFT-DO NOT QUOTE OR CITE
-------
1 Vose, J. M.; Swank, W. T. (1990) Preliminary estimates of foliar absorption of 15N-labeled nitric acid vapor (HNO3)
2 by mature eastern white pine (Pinus strobus). Can. J. For. Res. 20: 857-860.
3 Vuilleumier, L.; Harley, R. A.; Brown, N. J.; Slusser, J. R.; Kolinski, D.; Bigelow, D. S. (2001) Variability in
4 ultraviolet total optical depth during the Southern California Ozone Study (SCOS97). Atmos. Environ.
5 35: 1111-1122.
6 Wagrowski, D. M.; Hites, R. A. (1997) Polycyclic aromatic hydrocarbon accumulation in urban, suburban, and rural
7 vegetation. Environ. Sci. Technol. 31: 279-282.
8 Wall, D. H. (1999) Biodiversity and ecosystem functioning. BioScience 49: 107-108.
9 Wall, D. H.; Moore, J. C. (1999) Interactions underground: soil biodiversity, mutualism, and ecosystem processes.
10 Bioscience 49: 109-117.
11 Walton, J. R.; Johnson, J. B.; Wood, G. C. (1982) Atmospheric corrosion initiation by sulphur dioxide and
12 paniculate matter: II. characterisation and corrosivity of individual paniculate atmospheric pollutants.
13 Br. Corros. J. 17: 65-70.
14 Wang, W. C.; Yung, Y. L.; Lacis, A. A.; Mo, T.; Hansen, J. E. (1976) Greenhouse effects due to man-made
15 perturbations of trace gases. Science (Washington, DC) 194: 685-690.
16 Waring, R. H. (1987) Nitrate pollution: a particular danger to boreal and subalpine coniferous forests. In:
17 Fujimori, T.; Kimura, M., eds. Human impacts and management of mountain forests: [proceedings of a
18 symposium]. Ibaraki, Japan: Forestry and Forest Products Research Institute; pp. 93-105.
19 Waring, R. H.; Schlesinger, W. H. (1985) The carbon balance of trees. In: Forest ecosystems: concepts and
20 management. Orlando, FL: Academic Press, Inc.; pp. 7-37.
21 Watson, J. G.; Chow, J. C. (1994) Clear sky visibility as a challenge for society. Annu. Rev. Energy Environ.
22 19:241-266.
23 Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
24 (1990) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
25 Webb, A. H.; Bawden, R. J.; Busby, A. K.; Hopkins, J. N. (1992) Studies on the effects of air pollution on
26 limestone degradation in Great Britain. Atmos. Environ. Part B 26: 165-181.
27 Wedding, J. B.; Carlson, R. W.; Stukel, J. J.; Bazzaz, F. A. (1975) Aerosol deposition on plant leaves. Environ. Sci.
28 Technol. 9: 151-153.
29 Wedin, D. A.; Tilman, D. (1996) Influence of nitrogen loading and species composition on the carbon balance of
30 grasslands. Science 274: 1720-1723.
31 Weinbaum, S. A.; Neumann, P. M. (1977) Uptake and metabolism of 15N-labeled potassium nitrate by French prune
32 (Primus domestica L.) leaves and the effects of two surfactants. J. Am. Soc. Hortic. Sci. 102: 601-604.
33 Wellburn, A. R. (1990) Tansley review no. 24: Why are atmospheric oxides of nitrogen usually phytotoxic and not
34 alternative fertilizers? New Phytol. 115: 395-429.
35 Wells, B. W. (1939) A new forest climax: the salt spray climax of Smith Island, North Carolina. Torrey Bot. Club
36 Bull. 66: 629-634.
37 Wells, B. W.; Shunk, I. V. (1937) Seaside shrubs: wind forms vs. spray forms. Science 85: 499.
38 Wells, B. W.; Shunk, I. V. (1938) Salt spray: an important factor in coastal ecology. Torrey Bot. Club Bull.
39 65:485-492.
40 Wenny, B. N.; Schafer, J. S.; DeLuisi, J. J.; Saxena, V. K.; Barnard, W. F.; Petropavlovskikh, I. V.; Vergamini,
41 A. J. (1998) A study of regional aerosol radiative properties and effects on ultraviolet-B radiation.
42 J. Geophys. Res. [Atmos.] 103: 17,083-17,097.
43 Wesely, M. L. (1989) Parameterization of surface resistances to gaseous dry deposition in regional-scale numerical
44 models. Atmos. Environ. 23: 1293-1304.
45 Wesely, M. L.; Hicks, B. B. (1977) Some factors that affect the deposition rates of sulfur dioxide and similar gases
46 on vegetation. J. Air Pollut. Control Assoc. 27: 1110-1116.
47 Wesely, M. L.; Eastman, J. A.; Stedman, D. H.; Yalvac, E. D. (1982) An eddy-correlation measurement of NO2 flux
48 to vegetation and comparison to O3 flux. Atmos. Environ. 16: 815-820.
49 Wesely, M. L.; Cook, D. R.; Hart, R. L. (1983) Fluxes of gases and particles above a deciduous forest in wintertime.
50 Boundary Layer Meteorol. 27: 237-255.
51 Wesely, M. L.; Cook, D. R.; Hart, R. L.; Speer, R. E. (1985) Measurements and parameterization of paniculate
52 sulfur dry deposition over grass. J. Geophys. Res. [Atmos.] 90: 2131-2143.
53 Wesselink, L. G.; Meiwes, K.-J.; Matzner, E.; Stein, A. (1995) Long-term changes in water and soil chemistry in
54 spruce and beech forests, Soiling, Germany. Environ. Sci. Technol. 29: 51-58.
April 2002 4-240 DRAFT-DO NOT QUOTE OR CITE
-------
1 Westman, W. E. (1977) How much are nature's services worth? Measuring the social benefits of ecosystem
2 functioning is both controversial and illuminating. Science (Washington, DC) 197: 960-964.
3 Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
4 White, W. H. (1976) Reduction of visibility by sulphates in photochemical smog. Nature (London) 264: 735-736.
5 White, W. H. (1990) The components of atmospheric light extinction: a survey of ground-level budgets.
6 Atmos. Environ. Part A 24: 2673-2679.
7 White, W. H. (1991) Contributions to light extinction. In: Irving, P. M, ed. Acidic deposition: state of science and
8 technology, v. Ill: terrestrial, materials, health and visibility effects; report 24. Washington, DC: National
9 Acid Precipitation Assessment Program; pp. 85-102.
10 White, W. H.; Roberts, P. T. (1977) On the nature and origins of visibility-reducing aerosols in the Los Angeles air
11 basin. Atmos. Environ. 11: 803-812.
12 Williams, M. W.; Baron, J. S.; Caine, N.; Sommerfeld, R.; Sanford, R. (1996) Nitrogen saturation in the
13 Rocky Mountains. Environ. Sci. Technol. 30: 640-646.
14 Wilson, E. O. (1997) The creation of biodiversity. In: Raven, P. H., ed. Nature and human society: the quest for a
15 sustainable world: proceedings of the 1997 forum on biodiversity; October; sponsored by the National
16 Research Council. Washington, DC: National Academy Press.
17 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
18 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
19 Wiman, B. (1981) Aerosol collection by Scots pine seedlings: design and application of a wind tunnel method.
20 Oikos 36: 83-92.
21 Wiman, B. L. B. (1985) A non-stationary model for sulphur aerosol depletion and deposition in a young spruce
22 stand. Tellus Ser. B 37: 230-242.
23 Wiman, B. L. B.; Agren, G. I. (1985) Aerosol depletion and deposition in forests~a model analysis. Atmos.
24 Environ. 19: 335-347.
25 Wiman, B. L. B.; Lannefors, H. O. (1985) Aerosol characteristics in a mature coniferous forest-methodology,
26 composition, sources and spatial concentration variations. Atmos. Environ. 19: 349-362.
27 Wiman, B. L. B.; Agren, G. I.; Lannefors, H. O. (1985) Aerosol concentration profiles within a mature coniferous
28 forest—model versus field results. Atmos. Environ. 19: 363-367.
29 Winner, W. E.; Atkinson, C. J. (1986) Absorption of air pollution by plants, and consequences for growth.
30 Trends Ecol. Evol. 1: 15-18.
31 Winner, W. E.; Bewley, J. D. (1978a) Terrestrial mosses as bioindicators of SO2 pollution stress: synecological
32 analysis and the index of atmospheric purity. Oecologia 35: 221-230.
33 Winner, W. E.; Bewley, J. D. (1978b) Contrasts between bryophyte and vascular plant synecological responses in
34 an SO2-stressed white spruce association in central Alberta. Oecologia 33:311-325.
35 Witherspoon, J. P.; Taylor, F. G., Jr. (1971) Retention of 1-44 \i simulated fallout particles by soybean and
36 sorghum plants. HealthPhys. 21: 673-677.
37 Wittenburg, C.; Dannecker, W. (1992) Dry deposition and deposition velocity of airborne acidic species upon
38 different sandstones. In: Proceedings of the 1992 European aerosol conference; September; Oxford,
39 United Kingdom. J. Aerosol Sci. 23(suppl. 1): S869-S872.
40 Wolff, G. T. (1984) On the nature of nitrate in coarse continental aerosols. Atmos. Environ. 18: 977-981.
41 Wolff, G. T.; Collins, D. C.; Rodgers, W. R.; Verma, M. H.; Wong, C. A. (1990) Spotting of automotive finishes
42 from the interactions between dry deposition of crustal material and wet deposition of sulfate. J. Air Waste
43 Manage. Assoc. 40: 1638-1648.
44 Woodcock, A. H. (1953) Salt nuclei in marine air as a function of altitude and wind force. J. Meteorol. 10: 362-371.
45 World Health Organization. (1997) Nitrogen oxides. 2nd ed. Geneva, Switzerland: World Health Organization.
46 (Environmental health criteria 188).
47 World Meteorological Organization. (1988) Developing policies for responding to climatic change: a summary of
48 the discussions and recommendations of workshops; September-October 1987; Villach, Austria; and
49 November 1987; Bellagio, Austria. Geneva, Switzerland: World Meteorological Organization; report no.
50 WMO/TD; no. 225. [World Climate Impact Programme series report no. WCIP-1].
51 World Resources Institute. (2000) World resources 2000-2001: people and ecosystems: the fraying web of life.
52 Washington, DC: World Resources Institute.
53 Wu, P.-M.; Okada, K. (1994) Nature of coarse nitrate particles in the atmosphere~a single particle approach.
54 Atmos. Environ. 28: 2053-2060.
April 2002 4-241 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wyttenbach, A.; Tobler, L.; Bajo, S. (1989) Na, Cl and Br in needles of Norway spruce (P. abies) and in the aerosol
2 adhering to the needles. Toxicol. Environ. Chem. 19: 25-33.
3 Xu, J. R.; Balik, C. M. (1989) Infrared attenuated total reflectance study of latex paint films exposed to aqueous
4 sulfur dioxide. J. Appl. Polym. Sci. 38: 173-183.
5 Yanai, R. D.; Siccama, T. G.; Arthur, M.. A.; Federer, C. A.; Friedland, A. J. (1999) Accumulation and depletion
6 of base cations in forest floors in the northeastern United States. Ecology 80: 2774-2787.
7 Yarnel, B.; Kalkstein, L. S.; Scheraga, J. D. (2000) Mid-Atlantic regional assessment of climate change impacts.
8 dim. Res. (CR Special 7) 14: 153-269.
9 Yerrapragada, S. S.; Jaynes, J. H.; Chirra, S. R.; Gauri, K. L. (1994) Rate of weathering of marble due to dry
10 deposition of ambient sulfur and nitrogen dioxides. Anal. Chem. 66: 655-659.
11 Yocom, J. E.; Grappone, N. (1976) Effects of power plant emissions on materials. Palo Alto, CA: Electric Power
12 Research Institute; report no. EPRI/EC-139. Available from: NTIS, Springfield, VA; PB-257 539.
13 Yocom, J. E.; Upham, J. B. (1977) Effects of economic materials and structures. In: Stern, A. C., ed. Air pollution.
14 3rd ed. Volume II. New York, NY: Academic Press; pp. 93-94.
15 Young, P. (1996) Pollution-fueled "biodeterioration" threatens historic stone: researchers are probing the role
16 microorganisms play in the decay of buildings and artwork. Environ. Sci. Technol. 30: 206A-208A.
17 Young, J. R.; Ellis, E. C.; Hidy, G. M. (1988) Deposition of air-borne acidifiers in the western environment.
18 J. Environ. Qual. 17: 1-26.
19 Youngs, W. D.; Rutzke, M.; Gutenmann, W. H.; Lisk, D. J. (1993) Nickel and vanadium in foliage in the vicinity of
20 an oil-fired power plant. Chemosphere 27: 1269-1272.
21 Zappia, G.; Sabbioni, C.; Gobbi, G. (1993) Non-carbonate carbon content on black and white areas of damaged
22 stone monuments. Atmos. Environ. 27A: 1117-1121.
23 Zappia, G.; Sabbioni, C.; Pauri, M. G.; Gobbi, G. (1994) Mortar damage due to airborne sulfur compounds in a
24 simulation chamber. Mater. Struct. 27: 469-473.
25 Zappia, G., Sabbioni, C.; Riontino, C.; Gobbi, G.; Favoni, O. (1998) Exposure tests of building materials in urban
26 atmosphere. Sci. Total Environ. 224: 235-244.
27 Zdanovskii, A. B. (1948) New methods of calculating solubilities of electrolytes in multicomponent systems.
28 Zh. Fiz. Khim. 22: 1475-1485.
29 Zepp, R. G.; Callaghan, T. V.; Erickson, D. J. (1998) Effects of enhanced solar ultraviolet radiation on
30 biogeochemical cycles. J. Photochem. Photobiol. B 46: 69-82.
31 Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing characteristics and water content of
32 submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
33 27: 1593-1607.
34 Zipperer, W. C.; Wu, J.; Pouyat, R. V.; Pickett, S. T. A. (2000) The application of ecological principles to urban
35 and urbanizing landscapes. Ecol. Appl. 10: 685-688.
April 2002 4-242 DRAFT-DO NOT QUOTE OR CITE
-------
Appendix 4A
Colloquial and Latin Names
Alder, hazel
Alder, red
Bean, common
Beech
Birch, yellow
Blackhaw
Brush box
Ceanothus, hoaryleaf
Chaparral
Coach wood
Corn
Dogwood, flowering
Elm
Fir, balsam
Fir, Douglas
Fir, fraser
Grape
Grass, red brome
Grass, purple moor
Greenbriar
Gum, sweet
Haw, black
Heather, Scottish
Alnus serrulata (Aiton) Willdenow
Alnus rubra Bong.
Phaseolus vulgaris L.
Fagus sylvatica L.
Betula alleghaniensis Britt.
Viburnum prunifolium L.
Lophostemon confertus (R. BR.) P.G. Wilson & Waterhouse
Ceanothus crassifolius Torry
Ceanothus crassifolius
Ceratopetalum apetalum, D.Don
Zea mays L.
Cornus florida L.
Ulmus spp.
Abies balsamea (L) Mill.
Pseudotsuga menziesii (Mirb.) Franco.
Abies fraseri (Pursh.) Poir
Vitis spp.
Bromus rubens L.
Molina caerulea (L.) Moench.
Smilax spp.
Liquidambar styraciflua L.
Viburnum prunifolium 1.
Calluna vulgaris Salisb.
April 2002
4A-1
DRAFT-DO NOT QUOTE OR CITE
-------
Hickory
Hophornbeam
Ivy, English
Laurel, mountain
Lichen, monks hood
Maize
Maple, red
Maple
Maple, sugar
Mustard, small podded
Nettle, stinging
Oak, bur
Oak, English
Oak, chestnut
Oak, live
Oak, northern red
Oak, turkey
Oak, white
Oak
Oats, domestic
Oats, wild
Persimmon, common
Pine, eastern white
Pine, jack pine
Pine, loblolly
Pine, lodgepole
Pine, Scots (Scotch)
Pine, slash
Gary a spp.
Ostrya virginiana (Mill.) Koch
Hedera helix L.
Kalmia latifolia L.
Hypogymnia physiodes
Zea mays L.
Acer rubrum L.
Acer spp.
Acer saccharum Marsh.
Brassica geniculata L.
Urtica dioica L.
Quercus macrocarpa Michx.
Quercus rober L.
Quercusprinus = Q. montana Willd.
Quercus virginiana Mill.
Quercus rubra L.
Quercus laevis Walt.
Quercus alba L.
Quercus spp.
Avena sativa L.
Avenafatua L
Diosporos virginiana L.
Pinus strobus L.
Pinus banksiana Lamb.
Pinus taeda L.
Pinus contorta Loud.
Pinus sylvestris L.
Pinus elliotti Englem.
April 2002
4A-2
DRAFT-DO NOT QUOTE OR CITE
-------
Pine, lodgepole x jack pine
Poplar, black
Poplar, white
Poplar, yellow or tulip
Privet
Purple Moor Grass
Ragweed
Rhododendron, Catawba
Rhododendron, rosebay
Sage, coastal
Scottish Heather
Soybean
Spruce, Norway
Spruce, red
Spruce, sitka
Spruce, white
Sunflower
Sweetgum
Tulip poplar
Pinus contorta (Douglas ex Loud) x P. banksiana Lamb.
Populus nigra L.
Populus alba L.
Liriodendron tulipifera L.
Ligustrum spp.
Molina caerulea (L.) Moench.
Ambrosia spp.
Rhododendron catawbiense Michx.
Rhododendron maximum L.
Artemisia californica Less.
Calluna vulgar is Salisb.
Glycine max (L.) Merr.
Picea abies (L.) Karst.
Picea rubens Sarg.
Picea sitchensis (Bong.) Carr.
Picea glauca (Moench.) Voss.
Helianthus annuus L.
Liquidambar styraciflua L.
Liriodendron tulipifera L.
EUMYCOTA-FUNGI
Zygomycota
Vesicular Arbuscular Mycorrhizae
Scutellospora
Gigaspora
Glomus agrigatum
Glomus leptototicum
April 2002
4A-3
DRAFT-DO NOT QUOTE OR CITE
-------
Ascomycotina
Chaetomium sp.
Fungi Imperfect!
Aureobasidium pullulans
Cladosporium sp.
Epicoccum sp.
Pestalotiopsis
Phialophora verrucosa
Pleurophomella =Sirodothis
April 2002 4A-4 DRAFT-DO NOT QUOTE OR CITE
-------
i 5. HUMAN EXPOSURE TO PARTICULATE MATTER
2 AND ITS CONSTITUENTS
3
4
5 5.1 INTRODUCTION
6 5.1.1 Purpose
7 Exposure is defined as the contact by an individual with a pollutant for a specific duration
8 of time at a visible external boundary (modified from Duan 1982, 1991). For airborne particulate
9 matter (PM), the breathing zone is considered the point of contact; and the lung and heart are the
10 target organs of concern. An individual's exposure is measured as the PM air concentration in
11 his/her breathing zone over time. Understanding exposure is important, because it is the
12 individual who experiences adverse health effects associated with elevated PM concentrations.
13 Human exposure data and models provide the link between ambient monitoring data or
14 atmospheric models and lung deposition models to enable estimates of the source- air- exposure-
15 dose relationship for input into dose-response assessments for PM from ambient sources.
16 The goal of this chapter is to provide current information on the development of human
17 exposure data and models. This includes information on the relationships between PM measured
18 at ambient sites and personal exposures to PM from both ambient and nonambient sources and
19 the factors that effect these relationships. Human exposure data and models presented in this
20 chapter provide the critical link between ambient monitoring data, PM dosimetry, and
21 toxicological studies and epidemiological studies presented in other chapters. Specific objectives
22 of thi s chapter are fourfol d:
23 (1) To provide an overall conceptual framework of exposure science as applied to PM, including
24 the identification and evaluation of factors that determine personal PM exposure
25 (2) To provide a concise summary and review of recent data (since 1996) and findings from
26 pertinent PM exposure studies
27 (3) To characterize quantitative relationships between ambient air quality measurements (mass,
28 chemical components, number, etc.), as determined by a community monitoring site, and
29 total personal PM exposure, as well as its ambient and nonambient components
April 2002 5-1 DRAFT-DO NOT QUOTE OR CITE
-------
1 (4) To evaluate the implications of using ambient PM concentrations as a surrogate for personal
2 exposure in epidemiological studies of PM health effects
3 The U.S. Environmental Protection Agency's (EPA's) regulatory authority for PM applies
4 primarily to ambient air and those sources that contribute to ambient PM air concentrations.
5 Thus, a major emphasis must be to develop an understanding of exposure to PM from ambient
6 sources. However, personal exposure to total PM may result from exposure to PM from both
7 ambient and nonambient sources, and it is likely that both ambient and nonambient components
8 will have adverse health effects. Ultimately, it will be necessary to account for both in order to
9 fully understand the relationship between PM and health effects. In addition, an individual's
10 personal exposure to ambient, nonambient, and total PM would provide useful information for
11 studies where health outcomes are tracked individually.
12
13 5.1.2 Particulate Matter Mass and Constituents
14 Current EPA PM regulations are based on mass as a function of aerodynamic size.
15 However, EPA also measures the chemical composition of PM in both monitoring and research
16 studies. The composition of PM is variable and (as discussed in Chapters 7 and 8) adverse health
17 effects may be related to PM characteristics other than mass. Since PM from ambient and
18 nonambient sources also may have different physical and chemical characteristics, they may also
19 have different health effects. Ultimately, to understand and control health impacts caused by PM
20 exposures from all sources, it is important to quantify and understand exposure to those chemical
21 constituents responsible from various sources for the adverse health effects.
22 The National Research Council (NRC) recognized the distinction between measuring
23 exposure to PM mass and to chemical constituents when setting Research Priorities for Airborne
24 Particulate Matter I: Immediate Priorities and a Long-range Research Portfolio (NRC, 1998).
25 Specifically, NRC Research Topic 1 recommends evaluating the relationship between outdoor
26 measures versus actual human exposure for PM mass. The NRC Research Topic 2 recommends
27 evaluating exposures to biologically important constituents and specific characteristics of PM
28 that cause responses in potentially susceptible subpopulations and the general population. It also
29 was recognized by the NRC that "a more targeted set of studies under this research topic (#2)
30 should await a better understanding of the physical, chemical, and biological properties of
31 airborne particles associated with the reported mortality and morbidity outcomes" (NRC, 1999).
April 2002 5-2 DRAFT-DO NOT QUOTE OR CITE
-------
1 The NRC also stated that the later studies "should be designed to determine the extent to which
2 members of the population contact these biologically important constituents and size fraction of
3 concern in outdoor air, outdoor air that has penetrated indoors, and air pollutants generated
4 indoors" (NRC, 1999). Thus, exposure studies should include contributions from all sources.
5 The emphasis in this chapter on PM mass reflects the current state of the science. Where
6 available, data also have been provided on chemical constituents, although in most cases, the
7 data are limited. As recognized by the NRC, a better understanding of exposures to PM chemical
8 constituents from multiple sources will be required to more fully identify, understand, and
9 control those sources of PM with adverse health effects and to accurately define the relationship
10 between PM exposure and health outcomes due to either short-term or chronic exposures.
11
12 5.1.3 Relationship to Past Documents
13 Early versions of PM criteria documents did not emphasize total human exposure but rather
14 focused almost exclusively on outdoor air concentrations. For instance, the 1969 Air Quality
15 Criteria for Particulate Matter (National Air Pollution Control Administration, 1969) did not
16 discuss either exposure or indoor concentrations. The 1982 EPA PM Air Quality Criteria
17 Document (PM AQCD), however, provided some discussion of indoor PM concentrations,
18 reflecting an increase in microenvironmental and personal exposure studies (U.S. Environmental
19 Protection Agency, 1982). The new data indicated that personal activities, along with PM
20 generated by personal and indoor sources (e.g., cigarette smoking), could lead to high indoor
21 levels and high personal exposures to total PM. Some studies reported indoor concentrations that
22 exceeded PM concentrations found in the air outside the monitored microenvironments or at
23 nearby monitoring sites. Between 1982 and 1996, many more studies of personal and indoor PM
24 exposure demonstrated that, in most inhabited domestic environments, indoor PM concentrations
25 and personal PM exposures of the residents were greater than ambient PM concentrations
26 measured simultaneously (e.g., Sexton et al., 1984; Spengler et al., 1985; Clayton et al., 1993).
27 As a result, the NRC (1991) recognized the potential importance of indoor sources of
28 contaminants (including PM) in causing adverse health outcomes.
29 The 1996 AQCD (U.S. Environmental Protection Agency, 1996) reviewed the human PM
30 exposure literature through early 1996, mainly to evaluate the use of ambient monitors as
31 surrogates for PM exposure in epidemiology studies. Many of the studies cited showed poor
April 2002 5-3 DRAFT-DO NOT QUOTE OR CITE
-------
1 correlations between personal exposure or indoor measurements of PM and outdoor or ambient
2 site measurements. Conversely, Janssen et al. (1995) and Tamura et al. (1996a) showed that in
3 the absence of major nonambient sources, total PM exposures to individuals tracked through
4 time were highly correlated with ambient PM concentrations. Analyses of these latter two
5 studies led to consideration of ambient and nonambient exposures as separate components of
6 total personal exposure. As a result, the 1996 PM AQCD (U.S. Environmental Protection
7 Agency, 1996), for the first time, distinguished between ambient and nonambient PM personal
8 exposure. This chapter builds on the work of the 1996 PM AQCD by further evaluating the
9 ambient and nonambient components of PM, as well as reporting research that evaluates the
10 relationship between ambient concentrations and total, ambient, and nonambient personal
11 exposure.
12
13
14 5.2 STRUCTURE FOR THE CHAPTER
15 The chapter is organized to provide information on the principles of exposure, review the
16 existing literature, and summarize key findings and limitations in the information; the specific
17 sections are described below.
18 • Section 5.3 discusses the basic concepts of exposure, including definitions, methods for
19 estimating exposure, and methods for estimating ambient components of exposure.
20 • Section 5.4 presents PM mass data, including a description of the key available studies, the
21 relationship of PM exposures with ambient concentrations, and factors that affect the
22 relationship.
23 • Section 5.5 presents data on PM constituents, including a description of the key available
24 studies, the relationship with ambient concentrations, and factors that affect the relationship.
25 • Section 5.6 discusses the implications of using ambient PM concentrations in epidemiological
26 studies of PM health effects.
27 • Section 5.7 summarizes key findings and limitations of the information.
28
29
30
April 2002 5-4 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.3 BASIC CONCEPTS OF EXPOSURE
2 5.3.1 Components of Exposure
3 The total exposure of an individual over a discrete period of time includes exposures to
4 many different particles from various sources while in different microenvironments Cue's). Duan
5 (1982) defined a microenvironment as "a [portion] of air space with homogeneous pollutant
6 concentration." It also has been defined as a volume in space, for a specific time interval, during
7 which the variance of concentration within the volume is significantly less than the variance
8 between that microenvironment and surrounding ones (Mage, 1985). In general, people pass
9 through a series of microenvironments, including outdoor, in-vehicle, and indoor
10 microenvironments, as they go through time and space. Thus, total daily exposure for a single
11 individual to PM can be expressed as the sum of various exposures for the microenvironments
12 that the person occupies in the day (modified from National Research Council, 1991).
13 In a given microenvironment, particles may originate from a wide variety of sources. For
14 example, in an indoor microenvironment, PM may be generated by (1) indoor activities,
15 (2) outdoor PM entering indoors, (3) the chemical interaction of outdoor air pollutants and indoor
16 air or indoor sources, (4) transport from another indoor microenvironment, or (5) personal
17 activities. All of these disparate sources have to be accounted for when estimating total human
18 exposure to PM.
19 An analysis of personal exposure to PM mass (or constituent compounds) requires
20 definition and discussion of several classes of particles and exposure. In this chapter, PM
21 metrics may be described in terms of exposure or as an air concentration. PM also may be
22 described according to both its source (i.e., ambient, nonambient) and the microenvironment
23 where exposure occurs. Table 5-1 provides a summary of the terms used in this chapter, the
24 notation used for these terms, and their definition. These terms are used throughout this chapter
25 and provide the terminology for evaluating personal exposure to total PM and to PM from
26 ambient and nonambient sources.
27 The 1997 NAAQS were developed largely on the basis of evidence from epidemiological
28 studies that found relatively consistent associations between outdoor paniculate matter
29 concentrations and observed health effects. Thus an emphasis in this chapter is on
30 determinations of personal exposure to PM of ambient origin and the relationship between the
April 2002 5-5 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-1. CLASSES OF PARTICULATE MATTER EXPOSURE AND
CONCENTRATION DEFINITIONS
Term
Notation
Definition
Concentration
Personal Exposure
Microenvironment
Ambient PM
Ambient-Outdoor PM
Indoor PM
Ambient-Indoor PM
Primary Indoor-
generated PM
Secondary Indoor-
generated PM
Personal Exposure to
Indoor-Generated PM
Personal Exposure to
Indoor-Formed PM
Personal Exposure to
Personal-Activity PM
Personal Exposure to
Nonambient PM
C
Ca
Cao
Q
pig
^•do
Epact
-*-'nonag
General Definitions
Air concentration of PM in a given microenvironment, expressed in
Mg/rn3
Contact at visible external boundaries of an individual with a pollutant
for a specific duration of time; quantified by the amount of PM available
in concentration units Cwg/m3) at the oral/nasal contact boundary for a
specified time period (At ). General term for any exposure variable.
Volume in space, for a specific time interval, during which the variance
of concentration within the volume is significantly less than the variance
between that yx, and surrounding ,
Concentration Variables
PM in the atmosphere measured at a community ambient monitoring site
either emitted into the atmosphere directly (primary PM) or formed in it
(secondary PM). Major sources of PM species are industry, motor
vehicles, commerce, domestic emissions such as wood smoke, and
natural wind-blown dust or soil.
Ambient PM in an outdoor microenvironment
All PM found indoors
Ambient PM that has infiltrated indoors (i.e., has penetrated indoors and
remains suspended)
Primary PM generated indoors
Secondary PM generated by outdoor vapors reacting with indoor vapors
Exposure Variables
Eplg Sum of personal exposure resulting from primary indoor-generated PM
Sum of personal exposure resulting from secondary indoor-generated PM
Small-scale PM-generating activities that primarily influence exposure of
the person performing the activity itself
Sum of oersonal exposure to indoor-generated and personal activity PM
-^nonag ~~ ^sig ^sig ^pact
Personal Exposure to
Ambient-Generated PM
Personal Exposure to
Total PM
Sum of personal exposure caused by ambient-outdoor and ambient
indoor PM (does not include resuspended ambient PM previously
deposited indoors)
Sum of all personal exposures to ambient and nonambient PM
F = F +F +F +F=F +F
-M -^pig -^sig -^oact -^ag -^nonag -^ag
April 2002
5-6
DRAFT-DO NOT QUOTE OR CITE
-------
1 PM concentrations measured at ambient sites and personal exposure to PM. Although this is an
2 emphasis, it should be kept in mind that every particle that deposits in the lung becomes part of a
3 dose delivered to the individual. It is likely that the nonambient component of total exposure
4 also has health effects which would not be detected using community time-series epidemiology
5 studies. Since both ambient and nonambient components of PM exposure may have partial
6 influence on the ultimate dose and the health outcome, both components should be understood
7 and accounted for when assessing risk from PM and its constituents.
8
9 5.3.2 Methods To Estimate Personal Exposure
10 Personal exposure may be estimated using either direct or indirect approaches. Direct
11 approaches measure the contact of the person with the chemical concentration in the exposure
12 media over an identified period of time. Direct measurement methods include personal exposure
13 monitors (PEMs) for PM that are worn continuously by individuals as they encounter various
14 microenvironments and perform their daily activities. Indirect approaches use models and
15 available information on concentrations of chemicals in microenvironments, the time individuals
16 spend in those microenvironments, and personal PM generating activities to estimate personal
17 exposure. This section describes the methods to directly measure personal exposures and
18 microenvironmental concentrations, as well as the models used to estimate exposure. Several
19 approaches to estimate personal exposure to ambient PM also are described.
20
21 5.3.2.1 Direct Measurement Methods
22 5.3.2.1.1 Personal Exposure Monitoring Methods
23 In theory, personal exposure to total PM is measured by sampling the concentration of PM
24 in inhaled air entering the nose or mouth. Practically, it is defined as that PM collected by a
25 PEM worn by a person and sampling from a point near the breathing zone (but not impacted by
26 exhaled breath). PEMs for PM use measurement techniques similar to those used for ambient
27 PM. The PEM is a filter-based mass measurement of a particle size fraction (PM10 or PM2 5),
28 usually integrated over either a 24- or 12-h period at flow rates of 2 to 4 L/min using battery-
29 operated pumps. PEMs must be worn by study participants and, therefore, they must be quiet,
30 compact, and battery-operated. These requirements limit the type of pumps and the total sample
31 volume that can be collected. Generally, small sample volumes limit personal exposure
April 2002 5-7 DRAFT-DO NOT QUOTE OR CITE
-------
1 measurements to PM mass and a few elements detected by XRF. In most studies, PM25 and
2 PM10 have not been collected concurrently; thus, for personal exposure, there are very few data
3 available by which to estimate coarse thoracic PM (i.e., PM10_25).
4 Other methods used for ambient PM also have been adapted for use as a personal exposure
5 monitor. For example, a personal nephelometer that measures particle number within a specific
6 particle size range using light scattering has been used in personal exposure studies to obtain
7 real-time measurements of PM.
8
9 5.3.2.1.2 Microenvironmental Monitoring Methods
10 Direct measurements of microenvironmental PM concentrations, which are used with
11 models to estimate personal exposure to PM, also use methods similar to those for ambient PM.
12 These methods differ from PEMs in that they are stationary with respect to the microenvironment
13 (such as a stationary PEM). Microenvironmental monitoring methods include filter-based mass
14 measurements of particle size fractions (PM10, PM25), usually integrated over either a 24- or 12-h
15 period. Flow rates vary between various devices from 4 to 20 L/min. Larger sample volumes
16 allow more extensive chemical characterization to be conducted on microenvironmental samples.
17 Because more than one pumping system can be used in a microenvironment, PM25 and PM10 can
18 be collected simultaneously. Other continuous ambient PM measurement methods that have
19 been utilized for microenvironmental monitoring are the Tapered Element Oscillating
20 Microbalance (TEOM) and nephelometers. Various continuous techniques for counting particles
21 by size (Climet, LASX, SMPS, APS) also have also been used. Measurement techniques are
22 discussed in Chapter 2.
23
24 5 3.2.2 Indirect Methods (Modeling Methods)
25 5.3.2.2.1 Personal Exposure Models
26 Exposure modeling for PM mass (PM2 5 and PM10_2 5) and chemical constituents is a
27 relatively new field facing significant methodological challenges and input data limitations.
28 Exposure models typically use one of two general approaches: (1) a time-series approach that
29 estimates microenvironmental exposures sequentially as individuals go through time or (2) a
30 time-averaged approach that estimates microenvironmental exposures using average
31 microenvironmental concentrations and the total time spent in each microenvironment. Although
April 2002 5-8 DRAFT-DO NOT QUOTE OR CITE
-------
1 the time-series approach to modeling personal exposures provides the appropriate structure for
2 accurately estimating personal exposures (Esmen and Hall, 2000; Mihlan et al., 2000), a time-
3 averaged approach typically is used when the input data needed to support a time-series model
4 are not available. In addition, the time-varying dose profile of an exposed individual can be
5 modeled only by using the time-series approach (McCurdy, 1997,2000). We define the personal
6 exposure of an individual to a chemical in air to be (Lioy, 1990; NRC, 1991)
7
t —*2
E= \C(t)dt, (5-1)
8 where
9 E is the personal exposure during the time period from tt to t2, and
10 C(t) is the concentration near the nose and mouth not impacted by
11 exhaled air, at time t.
12 Even though the processes that lead to exposure are nonlinear in nature, personal exposure
13 models are often used to combine microenvironmental concentration data with human activity
14 pattern data in order to estimate personal exposures. Activity pattern data and information on
15 size, age, gender, and health status can be used to estimate inhalation rate. Time-averaged
16 models also can be used to estimate personal exposure for an individual or for a defined
17 population. Total personal exposure models estimate exposures for all of the different
18 microenvironments in which a person spends time, and total average personal exposure is
19 calculated from the sum of these microenvironmental exposures:
20
21 where Ey. is the personal exposure in each microenvironment,y (Duan, 1982). Example
22 microenvironments include outdoors, indoors at home, indoors at work, and in transit. Each
23 microenvironmental exposure, Ey, is calculated from the average concentration in
24 microenvironmenty, C , weighted by the time spent in microenvironmenty, t,.. T is the sum of ty
April 2002 5-9 DRAFT-DO NOT QUOTE OR CITE
-------
1 over ally. It is important to note that, although measurement data may be an average
2 concentration over some time period (i.e., 24 h), significant variations in PM concentrations can
3 occur during that time period. Thus, an error may be introduced if real-time concentrations are
4 highly variable, and an average concentration for a microenvironment is used to estimate
5 exposure when the individual is in that microenvironment for only a fraction of the total time.
6 This model has been applied to concentration data in a number of studies (Ott, 1984; Ott et al.,
7 1988, 1992; Miller et al., 1998; Klepeis et al., 1994; Lachenmyer and Hidy, 2000).
8 Microenvironmental concentrations used in the exposure models can be measured directly
9 or estimated from one or more microenvironmental models. Microenvironmental models vary in
10 complexity, from a simple indoor/outdoor ratio to a multi-compartmental mass-balance model.
11 A discussion of microenvironmental models is presented below in Section 5.3.2.2.2.
12 On the individual level, the time spent in the various microenvironments is obtained from
13 time/activity diaries that are completed by the individual. For population-based estimates, the
14 time spent in various microenvironments is obtained from human activity databases. Many of
15 the largest human activity databases have been consolidated by EPA's National Exposure
16 Research Laboratory (NERL) into one comprehensive database called the Consolidated Human
17 Activity Database (CHAD). CHAD contains over 22,000 person-days of 24-h activity data from
18 11 different human activity pattern studies (McCurdy et al., 2000). Population cohorts with
19 diverse characteristics can be constructed from the activity data in CHAD and used for exposure
20 analysis and modeling (McCurdy, 2000). These databases can also be used to estimate inhalation
21 rates based on activity levels, age, gender, and weight. Table 5-2 is a summary listing of the
22 human activity studies in CHAD.
23 Methodologically, personal exposure models can be divided into three general types:
24 (1) statistical models based on empirical data obtained from one or more personal monitoring
25 study, (2) simulation models based upon known or assumed physical relationships, and
26 (3) physical-stochastic models that include Monte Carlo or other techniques to explicitly address
27 variability and uncertainty in model structure and input data (Ryan, 1991; Macintosh et al.,
28 1995). The attributes, strengths, and weaknesses of these model types are discussed by Ryan
29 (1991), National Research Council (1991), Frey and Rhodes (1996), and Ramachandran and
30 Vincent (1999). A recent summary review of the logic of exposure modeling is found in Klepeis
31 (1999).
April 2002 5-10 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-2. ACTIVITY PATTERN STUDIES INCLUDED IN THE CONSOLIDATED HUMAN
ACTIVITY DATABASE (CHAD)
to
o
o
to
1
o
H
O
0
H
O
O
s
o
HH
H
W
Calendar
Study Time Period
Name of the Study Age1
Baltimore Jan-Feb 1997 65+
Jul-Aug 1998
CARB: Adolescents Oct 1987- 12 - 94
and Adults Sept 1988
CARB: Children Apr 1989- 0-11
Feb 1990
Cincinnati (EPRI) Mar-Apr and 0 - 86
Aug 1985
Denver (EPA) Nov 1982- 18-70
Feb 1983
Los Angeles: Elem. Oct 1989 10 - 12
School Children
Los Angeles: High Sept-Oct 1990 13 - 17
School Adoles.
National: NHAPS-A8 Sept 1992- 0 - 93
Oct 1994
National: NHAPS-B8 As above 0 - 93
University of Feb-Decl997 0-13
Michigan: Children
Valdez, AK Nov 1990- 11-71
Oct 1991
Washington, DC Nov 1982- 18-98
(EPA) Feb 1983
Diary
Days2 TvDe3 Time4
391 Diary; 15 -min 24h Standard
blocks
1,762 Retrospective 24h Standard
1,200 Retrospective 24h Standard
2,614 Diary 24h; nominal
7 p.m. -7 a.m.
805 Diary 24h; nominal
7 p.m.-7 a.m.
51 Diary 24h Standard
43 Diary 24h Standard
4,723 Retrospective 24h Standard
4,663 Retrospective 24h Standard
5,616 Retrospective 24h Standard
40 1 Retrospective Varying 24-h
period
699 Diary 24h; nominal
7 p.m. -7 a.m.
Notes: 'All studies included both genders. The age range depicted is for the subjects actually included; in most c
Age 0 = babies < 1 year old.
2The actual number of person-days of data in CHAD after the "flagging" and removal of questionable dati
'Retrospective: a "what did you do yesterday" type of survey; also known as an ex post survey. Diary: a "
"Standard = midnight-to-midnight.
'Was activity-specific breathing rate data collected?
6PEM = a personal monitoring study. In addition to the diary, a subject carried a small CO or PM2 5 moni
7P-D removed = The number of person-days of activity pattern data removed from consolidated CHAD bt
in the text.
8National Human Activity Pattern Study; A = the air version; B = the water version. The activity data ob1
9A question was asked regarding which activities (within each 6-h time block in the day) involved "heavy
Documentation
Rate5 or Reference Notes
No Williams et al. (2000a,b) Multiple days, varying from 5-15; part
of a PM25 PEM study
No Robinson etal. (1989)
Wiley etal. (1991a)
No Wiley etal. (1991b)
Yes Johnson (1989) 3 consecutive days; 1 86 P-D removed7
No Akland etal. (1985) Part of CO PEM6 study; 2 consec.
Johnson (1984) days; 5 5 P-D removed7
Yes Spier etal. (1992) 7 P-D removed7
Yes Spier et al. (1992) 23 P-D removed7
No9 Klepeis et al. (1995) A national random-probability survey
Tsang and Klepeis (1996)
No9 As above As above
No Institute for Social 2 days of data: one is a weekend day
Research (1997)
No Goldstein etal. (1992) 4 P-D removed7
No Akland et al. (1985) Part of a CO PEM6 study; 6 P-D
Hartwell et al. (1984) removed7
;ases, there was not an upper limit for the adult studies. Ages are inclusive.
\. See the text for a discussion of these procedures.
real-time" paper diary that a subject carried as he or she went through the day.
tor throughout the sampling period.
;cause of missing activity and location information; completeness criteria are listed
:ained on the two versions are identical.
• breathing", lifting heavy objects, and running hard.
-------
1 Personal exposure models that have been developed for PM are summarized in Table 5-3.
2 The regression-based models (Johnson et al., 2000; Janssen et al., 1997; Janssen et al., 1998a)
3 were developed for a specific purpose (i.e., to account for the observed difference between
4 personal exposure and microenvironmental measurements) and are based on data from a single
5 study, which limits their utility for broader purposes. Other types of models in Table 5-3 were
6 limited by a lack of data for the various model inputs. For example, ambient PM monitoring data
7 is not generally of adequate spatial and temporal resolution for these models. Lurmann and Korc
8 (1994) used site-specific coefficient of haze (COH) information to stochastically develop a time
9 series of 1-h PM10 data from every sixth day 24-h PM10 measurements. A mass-balance model
10 typically was used for indoor microenvironments when sufficient data were available, such as for
11 a residence. For most other microenvironments, indoor/outdoor ratios were used because of the
12 lack of data for a mass-balance model. In addition, only the deterministic model PMEX included
13 estimation of inhaled dose from activity-specific breathing rate information. Data from recent
14 PM personal exposure and microenvironmental measurement studies will help facilitate the
15 development of improved personal exposure models for PM.
16 An integrated human exposure source-to-dose modeling system that will include exposure
17 models to predict population exposures to environmental pollutants, such as PM, currently is
18 being developed by EPA/NERL. A first-generation population exposure model for PM, called
19 the Stochastic Human Exposure and Dose Simulation (SHEDS-PM) model, recently has been
20 developed. The SHEDS-PM model uses a 2-stage Monte Carlo sampling technique previously
21 applied by Macintosh et al. (1995) for benzene exposures. This technique allows for separate
22 characterization of variability and uncertainty in the model predictions (to predict the distribution
23 of total exposure to PM for the population of an urban/metropolitan area and to estimate the
24 contribution of ambient PM to total PM exposure). Results from a case study using data from
25 Philadelphia have been reported (Burke et al., 2001). Recently, the SHEDS model has been
26 extended by EOHSI scientists to provide estimates of integrated PM doses for different regions
27 of the lung for the Philadelphia case study population (Vyas et al, 2002). The inhalation model
28 uses dosimetry equations that account for anatomic, metabolic, and physical variability
29 information specified in the ICRP and HUMTRM models. These efforts are still preliminary but
30 critical for generating population based exposure and dose estimates by utilizing the available
31 dosimetric information and models described in Section 6. Ultimately, comprehensive
April 2002 5-12 DRAFT-DO NOT QUOTE OR CITE
-------
>
TABLE 5-3. PERSONAL EXPOSURE MODELS FOR PARTICIPATE MATTER
to
o
o
to
Study Citation
Model
Name
Model Type
Microenvironments or
Predictors
Output
Notes
Time-Series Models:
Hayes and PMEX
Marshall (1999)
Johnson et al.
(2000)
Klepeis et al.
(1994)
Lurmann and
Korc (1994)
Koontz and CPIEM
Niang (1998)
Deterministic Indoors: residential, work, school
Outdoors: near roadway, other
Motor vehicle
Auto travel, roadside, ETS, food
prep, grilling, high ambient PM
ETS, cooking, cleaning, attached
garage, wood burning
12 residential with different
sources, restaurant/bar,
nonresidential indoors, in transit,
outdoors
Stochastic Indoors: residence, office,
industrial plant, school, public
building, restaurant/lounge, other
Outdoors, in vehicle
Regression-based
Stochastic
REHEX-II Stochastic
Inhaled dose of PM10
Hourly for 24 h
By age/gender groups
Source contributions
PM2 5 exposure
24-h average
Respirable particle (PM3 5) exposure
Distribution of PM10 exposure for
population
Three averaging times (1 h, 24 h,
season)
Distribution of PM10 exposure for
population
Used IAQM
Used human activity data with activity-specific breathing rate
info.
Developed from scripted activity study (Chang et al., 2000)
Fixed I/O ratio of 0.7 for indoors w/o sources and 1.2 for in
transit
Reduced form mass balance model for indoors with PM
sources
Used California activity pattern and breathing rate data. Used
either a mass balance model or I/O ratio distribution for
indoor microenvironments. Indoor sources included.
Time-Averaged Models:
Clayton et al. SIM
(1999a)
Janssen et al.
(1997)
Janssen et al.
(1998a)
Ott et al. (2000) RCS
Burke et al.
(2001)
Stochastic
Regression-based
Regression-based
Statistical
SHEDS-PM Stochastic
Chao and Tung None
(2001)
Mass Balance
with Empirical
corrections
Smoking parent, ETS exposure,
outdoor physical activity
Number of cigarettes smoked,
hours of ETS exposure,
residence on busy road, time
in vehicle
Not separated
Outdoors, indoors: residence,
office, stores, school, in vehicle,
restaurant/lounge,
Indoors in unoccupied
residences in Hong Kong
Distribution of annual PM25
exposures
Accounts for difference between
personal and microenvironmental PM1(
Accounts for difference between
personal and microenvironmental
PM10
Distribution of PM10 exposure for
population
PM25 exposure distributions for
population, by age, gender, smoking
and employment status; PM2 5 exposure
uncertainty predictions. Percent
contribution from PM of ambient
origin to total personal exposures
Predictions of ambient PM in indoor
microenvironments
Based on 3-day ambient measurements
Children only
Adults only
A random-component superposition (RCS) model that uses
distribution of ambient PM10 and estimated nonambient PM10
concentrations.
Results for Ontario, Canada not corrected for 72-h compared
to 24-h averaging time in Riverside, CA and Phillipsburg, NJ.
A 2-stage Monte-Carlo simulation model for predicting
population distribution of daily- average personal exposures
to PM. Model has been applied to Philadelphia using spatially
and temporally interpolated PM2 5 ambient measurements
from 1992-1993 and 1990 census data. Does not consider
PM25 exposure from active smoking or exposure in subways.
Model makes corrections for nonideal mixing (residence with
multiple compartments with limited intermixing).
-------
1 evaluation of PM pollution and health data will utilize both exposure and dose metrics generated
2 for subgroups of concern.
3
4 5.3.2.2.2 Microenvironmental Models
5 The mass balance model has been used extensively in exposure analysis to estimate PM
6 concentrations in indoor microenvironments (Calder, 1957; Sexton and Ryan, 1988; Duan, 1982,
7 1991; McCurdy, 1995; Johnson, 1995; Klepeis et al., 1995; Dockery and Spengler, 1981; Ott,
8 1984; Ott et al., 1988, 1992, 2000; Miller et al., 1998; Mage et al., 1999; Wilson et al., 2000).
9 The mass balance model describes the infiltration of particles from outdoors into the indoor
10 microenvironment, the removal of particles in indoor microenvironments, and the generation of
1 1 particles from indoor sources:
12
13
V dC; /dt = vPCa -vQ-kVQ + Qi, (5-3)
14
15
16 where V = volume of the well-mixed indoor air (cubic meters),
17 Q = concentration of indoor PM;
18 v = volumetric air exchange rate between indoors and outdoors (cubic
19 meters per hour);
20 P = penetration ratio, the fraction of ambient (outdoor) PM that is not
21 removed from ambient air during its entry into the indoor volume;
22 Ca = concentration of PM in the ambient air (micrograms per cubic meter);
23 k = removal rate (per hour); and
24 Q; = indoor sources of particles (micrograms per hour).
25
26 Q; contains a variety of indoor, particle-generating sources, including: combustion or
27 mechanical processes; condensation of vapors formed by combustion or chemical reaction;
28 suspension from bulk material; and resuspension of previously deposited PM. The removal rate,
29 k, includes dry deposition to interior surfaces by diffusion, impaction, electrostatic forces, and
30 gravitational fallout. It may include other removal processes, such as filtration by forced air
3 1 heating, ventilation, or air-conditioning (HVAC), or by independent air cleaners. All parameters
April 2002 5-14 DRAFT-DO NOT QUOTE OR CITE
-------
1 except V are functions of time. P and k also are functions of particle aerodynamic diameter v,
2 and house characteristics. All variables in Equation 5-3 will have distributions within the
3 population and, in some cases, may change by a factor of 5 to 10. It is important to determine the
4 distribution of these variables. Sensitivity and uncertainty analyses are necessary when
5 attempting to explain the results.
6 In addition to the mass balance model, a number of single-source or single-
7 microenvironment models exist. However, most are used to estimate personal exposures to
8 environmental tobacco smoke (ETS). These models include both empirically based statistical
9 models and physical models based on first principles; some are time-averaged, whereas others
10 are time-series. These models evaluate the contribution of ETS to total PM exposure in an
11 enclosed microenvironment and can be applied as activity-specific components of total personal
12 exposure models. Examples of ETS-oriented personal exposure models are Klepeis (1999),
13 Klepeis et al. (1996, 2000), Mage and Ott (1996), Ott (1999), Ott et al. (1992, 1995), and
14 Robinson etal. (1994).
15
16 5.3.2.3 Methods for Estimating Personal Exposure to Ambient Particulate Matter
17 In keeping with the various components of PM exposure described above in Section 5.3.1,
18 personal exposure to PM can be expressed as the sum of exposure to particles from different
19 sources summed over all microenvironments in which exposure occurs. Total personal exposure
20 may be expressed as:
21
Et = Eag + Epig + Epact + Esig
(5-4)
Cjt — -t/ag ~T~ _t/nonag,
22
23 where Et is the total personal exposure to ambient and nonambient PM, Eag is personal exposure
24 to ambient PM (the sum of ambient PM while outdoors and ambient PM that has infiltrated
25 indoors, while indoors), Epig is personal exposure to indoor-generated PM, Epact is personal
26 exposure to PM from personal activity, Esig is exposure to indoor-formed PM, and Enonag is
27 personal exposure to nonambient PM. Again, this is a linear simplification of personal exposures
28 and ignores possible synergisms or interaction among indoor and outdoor pollutants. Although
April 2002 5-15 DRAFT-DO NOT QUOTE OR CITE
-------
1 personal exposure to ambient and nonambient PM cannot be measured directly, they can be
2 calculated or estimated from other measurement data. Approaches for estimating these
3 components of PM exposure are described in the following section.
4
5 5.3.2.3.1 Mass Balance Approach
6 Ambient-Indoor Concentrations of Particulate Matter
7 The mass balance model described above (Equation 5-3) has been used to estimate PM
8 concentrations in indoor microenvironments. This model also may be used to estimate ambient-
9 indoor (Cai) and indoor-generated (Cpig) PM concentrations. The mass balance model can be
10 solved for Cai and Cpig assuming equilibrium conditions, and assuming that all variables remain
11 constant (Ott et al., 2000; Dockery and Spengler, 1981; Koutrakis et al., 1992). By substituting
12 dCai + dCpig for dC; in Equation 5-3 and assuming dCai and dCpig = 0, ambient-indoor PM (Cai)
13 and indoor-generated PM (Cpig), at equilibrium, are given by
14
15 Cai=(Cao/Ptf)/(tf+k) (5-5)
(5-6)
16 where a = v/V, the number of air exchanges per hour. Equations 5-5 and 5-6 assume equilibrium
17 conditions and, therefore, are valid only when the parameters k, a, Cao, and Q; are not changing
18 rapidly and when the Cs are averaged over several hours. It should be understood that
19 equilibrium is a simplification of indoor microenvironments that are occupied by residents. This
20 assumption of equilibrium may only represent a virtual set of individuals or populations at risk.
21 Under certain conditions (e.g., air-conditioned homes, homes with HVAC or air cleaners that
22 cycle on and off, ambient pollutants with rapidly varying concentrations), nonequilibrium
23 versions of the mass balance model (Ott et al., 2000; Freijer and Bloeman, 2000; Isukapalli and
24 Georgopoulos, 2000) are likely to provide a more accurate estimate of Cai and Cpig. However, the
25 equilibrium model provides a useful, if simplified, example of the basic relationships (Ott et al.,
26 2000).
April 2002 5-16 DRAFT-DO NOT QUOTE OR CITE
-------
1 Equation 5-5 may be rearranged further to give Cai/Cao, the equilibrium fraction of ambient
2 PM that is found indoors, defined as the infiltration factor (FINF) (Dockery and Spengler, 1981).
3
F-^^ (5-7)
Cao a+k
4
5 The penetration ratio (P) and the decay rate (K) can be estimated using a variety techniques.
6 A discussion of these variables and estimation techniques is given in Section 5.4.3.2.2. Because
7 both P and k are a function of particle aerodynamic diameter, air exchange rate, and housing
8 characteristics, F^ also will be a function of these parameters. As a result F^ may present
9 substantial variability within a population. Distributions of this parameter should be estimated to
10 understand the uncertainty and variability associated with estimating exposure to PM of ambient
11 origin.
12
13 Personal Exposure to Ambient-Generated Particulate Matter
14 Personal exposure to ambient-generated PM (Eag) may be estimated using ambient-indoor
15 PM concentration (Cai) from the mass balance model, ambient outdoor PM concentrations (Cao)
16 and information on the time an individual spent in the various microenvironments.
17 Mathematically, this may be expressed as
18
(5-8)
(fl+k)
19
20 where y is the fraction of time that an individual spent outdoors, and (1 -y) is the fraction of time
21 spent indoors.
22 It is convenient to express personal exposure to ambient generated PM (Eag) as the product
23 of the ambient PM concentration (Cao or Ca) and a personal exposure or attenuation factor.
24 Following the usage in several recent papers (Zeger et al., 2000; Dominici et al., 2000; Ott et al.,
April 2002 5-17 DRAFT-DO NOT QUOTE OR CITE
-------
1 2000), the symbol a will be used for this attenuation factor. Equation 5-8 can be rearranged to
2 obtain an expression for a:
3
" Pa 1
C
ao
a + k
4
5 Substituting equation 5-7 in equation 5-9 gives a relationship for a in terms of the infiltration
6 factor FINF and the fraction of time spent in the various microenvironments:
7
a = y + (\-y)~FwF. (5-10)
8
9 Thus, personal exposures to ambient PM (Eag) may be calculated from measurable quantities:
10
11 Eag = «Cao. (5-11)
12
13 The factor a can be measured directly or calculated from measured or estimated values of the
14 parameters a, k, and P and the time spent in various microenvironments from activity pattern
15 diaries (Wilson et al., 2000). Since a depends on housing factors and lifestyle factors, air
16 exchange rate, and PM deposition rate, it could vary to a certain extent from region to region and
17 from season to season. Consequently, predicted exposures based on these physical modeling
18 concepts will provide exposure distributions derived conceptually as resulting from housing,
19 lifestyles, and meteorological considerations. For any given population the coefficient a may
20 represent substantial intra- and inter-personal variability, based on personal activities, housing
21 characteristics, particle size, and composition. Distributions of a should be determined using
22 population studies in order to evaluate the uncertainty and variability associated with model
23 exposures.
24 The use of a mass balance model to separate personal exposure into two components
25 because of exposure to ambient and nonambient concentrations is not novel. This approach,
26 based on Equation 5-4 as given in Duan (1982) and called superposition of component
April 2002 5-18 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations, has been applied using multiple microenvironments estimate exposures to carbon
2 monoxide (Ott, 1984; Ott et al., 1988, 1992), volatile organic compounds (Miller et al., 1998),
3 and particles (Koutrakis et al., 1992; Klepeis et al., 1994). However, in these studies, and in
4 most of the exposure literature, the ambient and nonambient components are added to yield a
5 personal exposure from all sources of the pollutant. The use of the mass balance model, ambient
6 concentrations, and exposure parameters to estimate exposure to ambient-generated PM and
7 exposure to indoor-generated PM separately as different classes of exposure has been discussed
8 in Wilson and Suh (1997) and in Wilson et al. (2000).
9
10 5.3.2.3.2 Tracer Species as Surrogates of Ambient-Generated Particulate Matter
11 The ratio of personal exposure to ambient concentration for a PM component that has no
12 indoor sources has often been used to calculate a (Wilson et al., 2000). Sulfate, in particular, is
13 often used as a marker of outdoor air in indoor microenvironments (Jones et al., 2000). It is
14 found primarily in the PM2 5 fraction of the aerosol (Cohen et al., 2000) and is formed in the
15 ambient air via photochemical oxidation of gaseous sulfur dioxide arising from the primary
16 emissions from the combustion of fossil fuels containing sulfur. It also arises from the direct
17 emissions of sulfur-containing particles from nonanthropogenic sources (e.g., volcanic activity,
18 wind-blown soil). In the indoor environment, the only common sources of sulfate may be
19 resuspension by human activity of deposited PM containing ammonium sulfates or soil sulfates
20 that were tracked into the home. In some homes an unvented kerosene heater using a high-sulfur
21 fuel may be a major contributor during winter (Leaderer et al., 1999). Use of matches to light
22 cigarettes or gas stoves can also be a source of sulfates. Studies that have used sulfate as a
23 surrogate for ambient PM are discussed in Section 5.4.3.1 (i.e., Oglesby et al., 2000; Sarnat et al.,
24 2000; Ebelt et al., 2000). When there are no indoor sources of fine-mode sulfates, one may
25 deduce that the ambient-to-personal relationship found for sulfates probably would be the same
26 as that for particulate matter of the same aerodynamic size range and physical/chemical
27 properties. This assumption has not been validated, however; and ambient PM with different
28 physical or chemical characteristics may not behave similarly to sulfate.
29
30
31
April 2002 5-19 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.3.2.3.3 Source-Apportionment Techniques
2 Source apportionment techniques provide a method for determining personal exposure to
3 PM from specific sources. If a sufficient number of samples are analyzed with sufficient
4 compositional detail, it is possible to use statistical techniques to derive source category
5 signatures, identify indoor and outdoor source categories, and estimate their contribution to
6 indoor and personal PM. Daily contributions from sources that have no indoor component can
7 be used as tracers to generate exposure to ambient PM of similar aerodynamic size or directly as
8 exposure surrogates in epidemiologic analyses. Studies that have used source-apportionment are
9 discussed in Section 5.4.3.3 (i.e., Ozkaynak and Thurston, 1987; Yakovleva et al., 1999; Mar
10 et al. 2000; Laden et al., 2000).
11
12
13 5.4 SUMMARY OF PARTICULATE MATTER MASS DATA
14 5.4.1 Types of Particulate Matter Exposure Measurement Studies
15 A variety of field measurement studies have been conducted to quantify personal exposure
16 to PM mass, to measure microenvironmental concentrations of PM, to evaluate relationships
17 between personal exposure to PM and PM air concentrations measured at ambient sites, and to
18 evaluate factors that affect exposure. In general, exposure measurement studies are of two types,
19 depending on how the participants are selected for the study. In a probability study, participants
20 are selected using a probability sampling design where every member of the defined population
21 has a known, positive probability of being included into the sample. Probability study results can
22 be used to make statistical inferences about the target population. In a purposeful or
23 nonprobability design, any convenient method may be used to enlist participants and the
24 probability of any individual in the population being included in the sample is unknown.
25 Participants in purposeful samples (also referred to as a "convenience" samples) may not have
26 same the characteristics that would lead to exposure as the rest of the unsampled population.
27 Thus, results of purposeful studies apply only to the subjects sampled on the days that they were
28 sampled and not to other periods of time. Although such studies may report significant
29 differences, confidence intervals, and/? values, they do not have inferential validity (Lessler and
30 Kalsbeek,1992). Purposeful studies may have generalizability (external validity). The extent of
April 2002 5-20 DRAFT-DO NOT QUOTE OR CITE
-------
1 generalizability is a matter of judgement based on study participant characteristics. Purposeful
2 studies of PM personal exposure can provide data to develop relationships on important exposure
3 factors and useful information for developing and evaluating either statistical or
4 physical/chemical human exposure models.
5 Regardless of the sampling design (probability or purposeful) there are two general
6 categories of study design that can be used to measure personal exposure to PM and evaluate the
7 relationship between personal PM exposure levels and ambient PM concentrations measured
8 simultaneously: longitudinal and cross-sectional. These are discussed in Section 5.4.3.1.1.
9
10 5.4.2 Available Data
11 5.4.2.1 Personal Exposure Data
12 Table 5-4 gives an overview of the personal exposure studies that have been conducted and
13 are reviewed in this section. This includes studies that have been reported since the 1996 PM
14 AQCD. Major studies that were reported before that time also have been included to provide a
15 comprehensive evaluation of data in this area. Table 5-4 gives information on the sampling and
16 study designs, the study population, the season, number of participants, PM exposure metric, and
17 the PM size fraction measured.
18 Although there are a number of studies listed in the table, the data available to evaluate
19 longitudinal relationships and the factors that influence these are limited. Few are based on
20 probability sampling designs that allow study results to be inferred to the general population and
21 to develop distributional data or exposures and the factors that affect exposure. Unfortunately,
22 none of these probability studies uses a longitudinal study design. This limits our ability to
23 provide population estimates and distributional data on the relationship between personal PM
24 exposures and ambient site measurements. In addition, most of the probability studies of PM
25 exposure were conducted during a single season; thus, variations in ambient concentrations, air
26 exchange rates, and personal activities are not accounted for across seasons. In these cases, study
27 results are only applicable to a specific time period. Longitudinal studies, on the other hand,
28 generally have small sample sizes and use a purposeful sampling design. Some studies did not
29 include ambient site measurements to allow comparisons with the exposure data, and
30 approximately half of these studies monitored PM2 5. Only one or two studies measured both
31 PM10 and PM2 5 to provide information on PM10_2 5.
April 2002 5-21 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-4. SUMMARY OF RECENT PM PERSONAL EXPOSURE STUDIES
^
to
o
to
H
6
o
0
H
O
_J
w
o
o
H
Study Design
Probability Studies
Pooled
Pooled
Pooled
Pooled
Purposeful Studies
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Study Location and
Population
Riverside, CA,
PTEAM
Basel, Switzerland,
EXPOLIS
Toronto, Canada
Mexico City
Wageningen,
Netherlands, school
children
Amsterdam (Am),
Helsinki (His), elderly
angina or coronary
heart disease
Baltimore, elderly
healthy and COPD
Fresno I
Fresno II (elderly)
Los Angeles, elderly
COPD subjects
Boston, COPD
subjects
Nashville, TN, COPD
subjects
Vancouver, British
Columbia, COPD
Amsterdam and
Wageningen, Neth.,
school children
No. of
Subjects
178
50
732
66
13
41 (Am)
49 (His)
21
5
16
30
18
10
16
45
Study Period
Fall 1990
1997
91 1995 -
8/1996
1992
1995
Winter 1998
Spring 1999
7-8/1998
Feb, 1999
Apr-May 1999
Summer/ Fall
1996
Winter 1996-7
Summer 1996
Summer 1995
April-Sept,
1998
1994, 1995
Days per PM Exposure"
Age Subject Metrics
10-70 l(12h) P, I, O, A
1 (48 h) P, I, P
16+ 3 P, I, O, A
< 65 1 P, I, O
10-12 6 P, A, School
50-84 22 (Am) P, I, O
27 (His)
72-93 5-22 P, I, O , A
60+ 24 P, I, O, A
24 P, I, O, A
56-83 4 P, I, O
12 P, I, O, A
36-88 6 P, I, O
54-86 7 P, A
10-12 4-8 P, A, School
Co-Pollutant
PM Size Measured1" Metrics Reference
PM10 Clayton etal. (1993)
Ozkaynak et al. (1996a,b)
PM2 5 VOC, CO, Oglesby et al. (2000)
NO2, S, K, Pb, Jantunen et al. (1998)
Br, Ca
PM2 5 (12 mo) Clayton et al. (1999a),
PM10(3mo) Pellizzari etal. (1999)
PM10 Santos-Burgoa et al. (1998)
PM25, PM10 Janssen et al. (1999a)
PM25 Janssen et al. (2000)
PM2 5, PM10 CO, O3, NO2, Williams et al. (2000a,b)
SO2
PM2 5, PM10 CO, O3 Evans et al. (2000)
PM2 5, PM10
PM25 Linn etal. (1999)
PM2 5 , PM10 Rojas-Bracho et al. (2000)
PM2 5 , PM10 Bahadori et al. (2001)
PM2 5, PM10 Ebelt et al. (2000)
PM10 Janssen etal. (1997)
-------
TABLE 5-4 (cont'd). SUMMARY OF RECENT PM PERSONAL EXPOSURE STUDIES (SINCE 1996)
Study Location and No. of
Co-Pollutant
to
o
o
to
^
to
OJ
o
§>
"T1
H
6
o
0
H
/O
O
w
o
o
HH
H
W
Study Design
131UU^ l^U^clllUll C111U
Population
1NU. Ul
Subjects Study Period Age
u&y^ pui
Subject
rivi UyA.pusuiu
Metrics
PM Size Measured11
\-Aj-r uiiuuui i
Metrics
Reference
Purposeful Studies (cont'd)
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Longitudinal
Pooled
Probability
Sample,
Pooled
Pooled
Longitudinal
Longitudinal
uE diary
Individual
Longitudinal
Pooled
Amsterdam, adults
Baltimore, elderly
subjects
Baltimore, elderly,
COPD, children
Tokyo, Japan,
elderly housewives
Osaka, Japan
Milan, Italy, office
workers
Indianapolis, IN
Banska Bystrica,
Slovakia
Wageningen, NL
Mpala, Kenya
Brunei
London, UK
Zurich, CH
37 1994 51-70
15 Summer 1998, 75 ± 6.8
Spring 1999
56 Summer 1998, Adults: 75±6.8
Winter 1999 Children: 9-13
COPD:
65±6.6
18 1992
26 Fall 1990-1995
100 Spring/ summer
and winter
240 1996 16 -?
49 1997-1998 15 -59
13 ?????? 12-14
252 1996-1998 5-75
5 1998 Adult
10 1997 9-11
10 1998 Adults
5-8
12
12
3
Multiple
days
1
One 72-h
sample/s
ubject
1
5-8
2 years
1
5 day/seas
on
3 seasons
12h/day
for
3 days
P, I, A
P
P, I, O, A
P, I, O, A
P, I, O
P, Home,
Office,
Commuting
P, I, A, O
P, I, O, A
P, A, I at school
I
P , PI, PO
P, I, O
P, I, O
PM10
PM2.5, PM10
PM2,
SPM
PM2,PM2.10,PM>10
PM10
PM2.5, PM10
PM10, PM2.5
PM2.5, PM10
Undefined
Optical MIE
Undefined
Optical MIE
PM2.5, PM10
Pollen
O3, NO2, SO2
VOCs
O3, NO2, SO2,
CO, EC,/OC.
VOC
NO2
NO2, CO,
VOCs
Mn, Al, Ca
SO4=, nicotine
None
CO
None
None
None
Janssenetal. (1998a)
Sarnat et al. (2000)
Sarnat et al. (2000)
Tamuraetal. (1996a)
Tamuraetal. (1996b)
Carreretal. (1998)
Pellizzari et al. (2001)
Brauer et al. (2000)
Janssenetal. (1999a)
Ezzati and Kammen (20
01)
Muraleedharan TR,
Radojevic M (2000).
Wheeler et al. 2000
Riediker et al. 2000
"All based on gravimetric measurements.
bP = personal, I = indoors, O = outdoors, A = ambient.
-------
1 Four large-scale probability studies that quantify personal exposure to PM under normal
2 ambient source conditions have been reported in the literature. These include the EPA's Particle
3 Total Exposure Assessment Methodology (PTEAM) study (Clayton et al., 1993; Ozkaynak et al.,
4 1996a,b); the Toronto, Ontario, study (Clayton et al., 1999a and Pellizzari et al., 1999); the Air
5 Pollution Exposure Distribution within Adult Urban Populations in Europe (EXPOLIS) exposure
6 study (Jantunen et al., 1998, 2000; Oglesby, et al., 2000); and a study of a small, highly polluted,
7 area in Mexico City (Santos-Burgoa et al., 1998). Only preliminary results have been reported
8 for the EXPOLIS study. A fifth study conducted in Kuwait during the last days of the oil-well
9 fires (Al-Raheem et al., 2000) is not reported here because the ambient PM levels were not
10 representative of normal ambient source conditions.
11 Recent longitudinal exposure studies have focused on potentially susceptible
12 subpopulations such as the elderly with preexisting respiratory and heart diseases (hypertension,
13 chronic obstructive pulmonary disease, and congestive heart disease). This is in keeping with
14 epidemiologic studies that indicate mortality associated with high levels of ambient PM2 5 is
15 greatest for elderly people with cardiopulmonary disease (U.S. Environmental Protection
16 Agency, 1996). Longitudinal studies were conducted in the Netherlands by Janssen (1998) and
17 Janssen et al. (1997, 1998a,b, 1999b,c) on purposefully selected samples of adults (50 to 70 years
18 old) and children (10 to 12 years old). Several additional studies have focused on nonsmoking
19 elderly populations in Amsterdam and Helsinki (Janssen et al., 2000), Tokyo (Tamura et al.,
20 1996a), Baltimore (Liao et al., 1999; Williams et al., 2000a,b,c), and Fresno, CA (Evans et al.
21 2000). These cohorts were selected because of the low incidence of indoor sources of PM (such
22 as combustion or cooking). This should allow an examination of the relationship between
23 personal and ambient PM concentrations without the large influences caused by smoking,
24 cooking, and other indoor particle-generating activities. The EPA has a research program
25 focused on understanding PM exposure characteristics and relationships. Within the program,
26 longitudinal studies are being conducted on elderly participants with underlying heart and lung
27 disease (COPD, patients with cardiac defibrillator, and myocardial infarction), an elderly
28 environmental justice cohort, and asthmatics. These studies are being conducted in several cities
29 throughout the United States and over several seasons. Only preliminary data are currently
30 available, and results are not reported in this document.
April 2002 5-24 DRAFT-DO NOT QUOTE OR CITE
-------
1 A series of studies by Phillips et al. (1994, 1996, 1997a,b, 1998a,b, 1999) examined
2 personal ETS exposure in several European cities. Participants varied by age and occupation.
3 Respirable Paniculate Matter (RSP) concentrations were reported. These studies are not
4 included in Table 5-4, because of their focus on ETS exposure (which is not the focus of this
5 chapter). A small personal exposure study in Zurich, Switzerland, was reported by Monn et al.,
6 (1997) for PM10. This study also is not listed in Table 5-4, because indoor and outdoor
7 measurements were not taken simultaneously with the personal measurements and other details
8 of the study were not published.
9
10 5.4.2.2 Microenvironmental Data
11 Usually, personal PM monitoring is conducted using integrated measurements over a 12- or
12 24-h period. As such, total PM exposure estimates based on PEM measurements do not capture
13 data from individual microenvironments. Recent studies have examined PM concentrations in
14 various microenvironments using a number of different types of instruments ranging from filter-
15 based to continuous particle monitors. Details on the instruments used, measurements collected,
16 and findings of these studies according to microenvironment (residential indoor, nonresidential
17 indoor, and traffic-related) are summarized in Table 5-5. Those studies which collected
18 microenvironmental data as part of a personal exposure monitoring study are summarized in
19 Table 5-4. In general, the studies listed in Table 5-5 are relatively small, purposeful studies
20 designed to provide specific data on the factors that effect microenvironmental concentration of
21 PM from both ambient and nonambient sources.
22 Recently published studies have used various types of continuous monitors to examine
23 particle concentrations in specific microenvironments and resulting from specific activities.
24 Continuous particle monitors such as the SMPS, APS, and Climet have been used to measure
25 particle size distributions in residential microenvironments (Abt et al., 2000a; Long et al., 2000a;
26 Wallace et al., 1997; Wallace, 2000a; McBride et al., 1999; Vette et al., 2001). These studies
27 have been able to assess penetration efficiency for ambient particles to indoor
28 microenvironments, as well as penetration factors and deposition rates. Continuous instruments
29 are also a valuable tool for assessing the impact of particle resuspension caused by human
30 activity. A semi-quantitative estimate of PM exposure can be obtained using personal
31 nephelometers that measure PM using light-scattering techniques. Recent PM exposure studies
April 2002 5-25 DRAFT-DO NOT QUOTE OR CITE
-------
>
TABLE 5-5. SUMMARY OF RECENT MICROENVIRONMENTAL PM MEASUREMENT STUDIES
to
o
o
to
Reference
Study
Description
Size Fraction
Instrument(s)
Summary of Measurements
Notes/Findings
to
Residential Indoor: Nonsmoking Homes
Abt et al. (2000a)
Boston, MA
Long et al. (2000a)
Boston, MA
2 homes,
2 seasons,
6 days
9 homes,
2 seasons
SMPS
APS
SMPS
TEOM
Detailed indoor/outdoor traces of PM in
various size classes for different air
exchange rates (< Ih "' to > 2 h"1)
0.02-10 Continuous PM distributions and size
distributions obtained for indoor and
outdoor air
Major indoor sources of PM: cooking, cleaning, human
activity
Sources of fine particles: cooking and outdoor particles;
Sources of coarse particles: cooking, cleaning, indoor
activities. 50% of particles by volume generated by
Anuszewski et al. (1998)
Seattle, WA
Leaderer et al. (1999)
Southwest, VA
9 homes,
18 days
58 homes, summer
Wallace etal. (1997),
Wallace (2000b)
Reston, VA
Howard-Reed et al. (2000)
Fresno, CA
Baltimore, MD
Rea etal. (2001)
Baltimore, MD
Fresno, CA
1 home, 4 years
15 participants
15 participants
Nephelometer 2.5
(radiance)
10
SMPS
Climet
PAHs
Black carbon
6 size bins;
100 size
channels
0.01-
0.4 ,um
Nephelometer 0.1-10
(personal
MIE) 2.5
PEM
Nephelometer 0.1-10
(personal
MIE) 2.5 and 10
PEM
Simultaneous indoor and outdoor PM
measured continuously; 1-h avg time,
I/O = 0.98; air exchange rate: 0.7-1.7 h'1
24 h mean: Regional air 26.0 ± 11.5
,ug/m3 (n = 47); Outdoor homes 28.0 ±
17.7 //g/m3 (n = 43); Indoor w/ AC 28.9
± 18.7 //g/m3 (n = 49); Indoor w/o AC
33.3 ± 14.2,ug/m3(n = 8)
24 h mean: Regional air 20.2 ± 9.9 //g/m3
(n = 50); Outdoor homes 21.8 ± 14.8
Mg/m3 (n = 43); Indoor w/ AC 18.7 ±
13.2 //g/m3 (n = 49); Indoor w/o AC 21.1
±7.5 /ig/m3 (n = 9)
Time activity data, whole-house air
exchange rates
Continuous carbon monoxide: descriptive
data for monitored pollutants; size
profiles for six indoor particle sources
Continuous (15-min avg) PM and time
activity data; 24-h PM mass; participants
from Baltimore and Fresno PM panel
studies.
Descriptive statistics from each study for
five microenvironments
Continuous (15-min avg) PM and time
activity data; 24-h PM mass;
Modeled PM mass and time activity data
to apportion time spent in a location.
Good comparison with nephelometer
mass (6-20%)
indoor events were ultrafine particles.
Homes contained asthmatic children, heavy wood
burning.
Dominant source of fine particles was outdoor air.
Epidemiological study of maternal and infant health
effects associated with indoor air pollution
Indoor PM concentrations were lower for homes with air
conditioning (AC) than non-air-conditioned homes.
0.3- to 0.5-/an particles linked to outdoor concentrations,
frying, broiling; 0.5- to 2.5-^m particles related to
cooking events; >2.5-^m particles influenced by physical
movement.
Time-series plots of personal nephelometer data showed
that each participant's PM exposure consisted of a series
of short-term peaks, imposed on a background caused by
ambient PM concentrations.
54 ± 31% of average daily PM2 5 exposure occurred
indoor residences, where participants spent 83 ± 10% of
their time. A significant portion of PM2 5 exposure
occurred where participants spent 4-13% of their time.
-------
5-5(cont'd). SUMMARY
MICROENVIRONMENTAL PM
STUDIES
T3
to
O
to
Reference
Residential Indoor:
Quintana et al. (2000)
San Diego, CA
Chang et al. (2000)
Baltimore, MD
Study
Description
Nonsmoking Homes (cont'd)
Asthmatic children
indoor and outdoor
9 homes
1 person
performing
predetermined
activities
Instrument(s)
Nephelometer
(personal
MIE)
Harvard
impactors
TEOM
"Roll around"
monitor
(RAS)(PM25,
CO, VOC, O3
NO2 SO2)
Size Fraction Summary of Measurements
0.1-10 Indoor and outdoor measurements
collected using passive, active, and active
2.5 and 10 heated nephelometers for comparison to
PM mass measurements.
2.5 1-h personal exposures measured
simultaneously. Personal and ambient
concentrations were compared.
Notes/Findings
Nephelometer correlates best with PM2 5
vs. Indoor PM25 r = 0.66 vs. indoor PM10 r = 0.13
vs. outdoor PM25 r = 0.42 vs. outdoor PM10 r = 0.20
1-h personal O3 exposures were significantly lower in
indoor than outdoor microenvironments.
1-h personal CO exposures were highest in vehicles.
to
Lioyetal. (1999)
NA
10 vacuum cleaners
0.3-0.5 Vacuum cleaners ranged in collection
efficiency from 29-99%.
Personal and ambient PM2 5 correlations were strongest
for outdoor microenvironments and those with high air
exchange rates (i.e., vehicles).
Substantial fine particle emissions from motors with
emission rates from 0.028 - 128.8 /^g/min.
Ezzati and Kammen (2001)
Mpala , Kenya
O
w
£;
^ Chao and Tung (200 1 )
D Hong Kong
O
2|
O
H
0
O Fischer et al. (2000)
H Amsterdam, NL
rn
o
73
o
H
w
55 Native Huts MiniRam
2- years (MIE)
5 unoccupied Dust-Trak
homes measured (TSI)
indoors and
outdoors, along
with air exchange
rates
Measured traffic Harvard
related differences Impactors
ofPMandVOCs,
indoor/outdoor in
18 paired homes at
varying distances
from traffic.
Not specified.
Optical
Device
detects
particles 1-10
um, but it is
not PM10
PM2 5 real
time,
calibrated
against an
Andersen
Mark II
PM25and
PM10. EC
was measured
by reflectance
ofthePM25
filters. PAH
also measured
as indicator of
Diesel traffic
Measured PM surrounding wood fires in
un vented huts. PM measures were up to
8000 ug/m3, but uncalibrated against
wood smoke
In the limit as air exchange goes to zero,
there appears to be a r esidual source,
perhaps from drafts or thermal effects.
Above air exchange rates of 4.5/hr
penetration goes to 1, but indoor
turbulence resuspends previously settled
PM25
Outdoor PM10 and PM2 5 were
approximately 15-20% higher at higher
traffic streets than at the quiet streets on
the same days. However, much larger
differences were found for PAH and EC
which are traffic specific
Exposures were related to ARI
Developed an excellent model for ambient PM infiltration
in the absence of anthropogenic indoor sources.
"This [study] supports the use [of] traffic related
pollution mapping as an exposure proxy in large-scale
epidemiologic studies into health effects of motorized
traffic emissions."
-------
5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL PM
STUDIES
T3
to
O
O
to
Study
Reference Description
Residential Indoor: Nonsmoking Homes (cont'd)
Kingham et al. (2000) Measured PM at
Huddersfield, UK ten homes of non-
smokers, < 50 m
and
>300m from
traffic.
Morawska et al. (2001) Measured PM
Brisbane, Australia indoors and
outdoors at
16 homes while
residents were
absent. Air
exchange rate
estimated, not
measured.
Instrument(s)
Harvard
Impactors
Scanning
mobility
particle sizer,
aerodynamic
particle sizer,
and a TSI
dust-trak
Size Fraction
PM2 5 and
PM10 and
PAH. EC
measured by
filter
reflectance .
Submicron
PM,
Supramicron
PM,
PM2.5
Summary of Measurements
Median Indoor/outdoor ratio =1
(no indoor combustion sources)
For supra and sub micron particles,
indoor = outdoor for normal ventilation
conditions of > 2 air changes/hour.
Notes/Findings
Found an absence of a spatial gradient, perhaps due to
wind direction effects (e.g. sometimes upwind and
sometimes downwind of traffic)
Average outdoor PM concentrations are good estimates of
average indoor concentrations of PM of ambient origin
for air exchange rates of > 0.5/h
Residential Indoor: Other Home Types
l^J
oo
o
J>
H
1
O
o
2|
O
H
0
^
0
H
O
H
W
Brauer etal. (1996)
Mexico
Jenkins etal. (1996 a,b)
16 U.S. Cities
McBride etal. (1999)
NA
Vette etal. (2001)
Fresno, CA
22 rural Mexican
homes
(smoking and
nonsmoking)
Smoking and
nonsmoking homes
Combustion source
(incense) and
walking
(1 room, carpeted)
Detached
semioccupied
residence
Inertial 10
impactor
Nephelometer 2.5
(Radiance)
Fluoropore Particle phase
membrane ETS markers
filters
Met-One laser
particle
counter
SMPS 0.01-2.5
LASX
Indoor PM2.5: 132-555 ,ug/m3 PM10:
282-768 |/g/m3. Outdoor PM25: 37 ,ug/m3
PM10: 68 Mg/m3; I/O PM25: 1.8-12.4;
PM10: 4.7-10.0
Mean PM3 5 concentrations were 17-20
//g/m3 in smoking homes over
nonsmoking homes
Ratios of particle counts a 1.0 and 5.7 m
from the combustion source/activity were
obtained
Temporal relationships between indoor
and outdoor aerosol concentrations
evaluated; penetration factors and
deposition rates estimated. Fresno panel
study empty residence
Variety of cooking fuels used
Nephelometer data were highly correlated with PM2 5
PM10 indoors (r = ~0. 87-0.95)
and
Proximity to source may help explain the existence of a
personal cloud.
Diurnally variable indoor/outdoor aerosol concentration
ratios because of resuspension from daytime activities.
Penetration factors ranged from 0.5 to 0.9.
-------
5-5 (cont'd). SUMMARY OF RECENT MICROENVIRONMENTAL PM MEASUREMENT STUDIES
to
o
to
1
to
2
s>
'Tl
H
6
o
^
O
H
0
0
H
W
0
^
o
H
w
Reference
Study
Description Instrument(s)
Size Fraction Summary of Measurements
Notes/Findings
Nonresidential Microenvlroranents
Bohadana et al. (2000)
Donham et al. (2000)
San Francisco, CA
Klepeisetal. (1996)
San Francisco, CA
Nieuwenhuij sen et al. (1999)
Teschkeetal. (1999)
Baeketal. (1997)
Korea
Ottetal. (1996)
California
Houseman etal. (2001)
Boston, MA
Brauer and Mannetje (1998)
Vancouver, BC
Lee and Chang (1999)
Hong Kong
Manufacturing
plant, woodworkers
34 poultry workers NIOSH
Method 0600
monitors
probed
respirators
Airport lounge, TSI8510
ETS piezobalance
Agricultural
activities
Wood production
wood finishing
wood construction
workers
Indoor and outdoor
smoking
restaurants
Bar before and after Piezobalance
smoking prohibited
Indoor and outdoor TSI
restaurants, DUSTRAK
stores
indoor restaurants,
various smoking
policies
indoor and outdoor
5 classrooms
Not given 443 personal time-weighted average
occupations samples of airborne dust
5 Total dust sampled indoor respiratory
masks.
Personal monitoring: 630 ± 980 Mg/m3
(n = 210) ranging from 10-7,730 Mg/m3
3.5 Estimated cigarette emission rate of
1.43 mg/min/cigarette.
4 Average respirable fraction: 4.5 mg/m3
~50 1,632 observations from 1979-1997.
Arithmetic mean exposure: 7.93 mg/m3
Geometric mean exposure: 1.86 mg/m3
3.5 Indoor concentrations: 33-475 Mg/m3
Outdoor concentrations: 12-172 Mg/m3
I/O: 2.4
3.5 Smoking permitted:
indoor 26.3-182 Mg/m3; outdoor <5-67
Mg/m3
Smoking prohibited:
indoor 4-82 Mg/m3; outdoor 2-67 Mg/m3
10 Indoor restaurants: 14-278 Mg/m3
Outdoor restaurants: 7-281 Mg/m3
Indoor stores: 12-206 Mg/m3
Outdoor stores: 7-281 Mg/m3
2.5 Nonsmoking: PM2 5 7-65 Mg/m3;
10 PM10 <10-74 Mg/m3
Restricted smoking (>40% nonsmoking)
PM2.5 11-163 Mg/m3 ; PM10 24-89 Mg/m3
Unrestricted smoking: PM25 47-253
Mg/m3 ;PM10 51-268 Mg/m3
10 Indoor PM10: 30-470 Mg/m3
Outdoor PM10: 20-617 Mg/m3
Respirable dust constituted about W,
measured.
'o of total dust
Personal exposures to ETS can be modeled in these types
of microenvironments .
No significant correlation between indoor and outdoor
measurements.
I/O nonsmoking: 2.2.
I/O smoking: 3.4
Avg I/O for restaurants: 2.3
Not known if the restaurants allowed
smoking
In stores, indoor and outdoor measurements were
correlated, avg I/O: 0.83
-------
5-5(cont'd). SUMMARY OF
MICROENVIRONMENTAL PM
STUDIES
T3
to
O
to
Reference
Study
Description
Instrument(s)
Size Fraction Summary of Measurements
Notes/Findings
Traffic-Related Microeiwironments
Praml and Schierl (2000)
Munich, Germany
Trams and buses,
rural and urban
Continuous
millipore
polycarbonate
filter
10 n = 201 4-h trips, mean concentration
155 ,ug/m3 range: 13-686 //g/m3
I/O: 2.8
Tram > circular bus route > radial bus route
Day > night
5: greatest concentrations by
bicycling and buses
Background and 2.5 PM25 background: 21-35 Mg/m3; roadway
roadway 10 23-43 //g/m3
PM10 background: 13-32 and 29-62
//g/m3;
roadway 16-56 and 30-75 //g/m3
Morning commutes were generally higher than afternoon
commutes; relationships determined between PM and
wind speed and vehicle speed
MeanPM1027.3±3.0,ug/m3
No vertical gradient (0-20 m) and horizontal gradient
(0-80 m) in distance from road, each about 13%.
No significant differences between wet and dry periods.
Vehicles in front of the monitored vehicle accounted for
most of the in-vehicle commuting exposure; average I/O:
0.6-0.8 h"1 for PM25; carpool lane concentrations were
30-60% lower than noncarpool lane concentrations
No concentration gradient with increasing distance from
the roadways for PM2 5 and PM10; concentration gradient
did exist for black smoke, also found an effect with wind
direction
The average in-vehicle to outdoor ratio was 0.99.
Average I/O: 3; subway values were correlated with
outdoor concentrations.
Average roadway/background ratio: 3 for PM2 5 and PM10.
Average increase in concentration at the roadway
7.2-12.7
-------
1 have used personal nephelometers (1 min avg time) to measure PM continuously (Howard-Reed
2 et al., 2000; Quintana et al., 2000) in various microenvironments. These data have been used to
3 identify the most important ambient and nonambient sources of PM, to provide an estimate of
4 source strength, and to compare modeled time activity data and PEM 24-h mass data to
5 nephelometer measurements (Rea et al., 2001). Several studies also have examined PM exposure
6 in vehicles using both continuous and filter-based techniques.
7
8 5.4.2.3 Reanalyses of Previously-Reported Particulate Matter Exposure Data
9 Several papers that have been published recently that reanalyzed and interpreted the data
10 collected in previous PM exposure studies are summarized in Table 5-6. These reanalyses are
11 directed towards understanding the personal cloud, the variability in total PM exposure, and the
12 personal exposure-to-ambient concentration relationships for PM. Results of these reanalyses are
13 summarized in Table 5-6 and given in more detail in Section 5.4.3. Brown and Paxton (1998)
14 determined that the high variability in personal exposure to PM makes the personal-to-ambient
15 PM relationship difficult to predict. Wallace (2000b) used data from a number of studies to test
16 two hypotheses: elderly COPD patients have (1) smaller personal clouds and (2) higher
17 correlations between personal exposure and ambient concentrations, compared to healthy elderly,
18 children, and the general population. The analysis by Wallace (2000a) and three subsequent
19 longitudinal studies (Williams 2000a,b,c; Ebelt et al., 2000; Sarnat et al., 2000) support
20 hypothesis 1 but not hypothesis 2. Ozkaynak and Spengler (1996) show that at least 50% of
21 personal PM10 exposure for the general population is because of ambient particles that
22 significantly contribute to inhaled particles. Wilson and Suh (1997) conclude that fine and
23 coarse particles should be treated as separate classes of pollutants because of differences in
24 characteristics and potential health effects. Wilson et al. (2000) give a review of what they call
25 the "exposure paradox" and determine that personal PM needs to be divided into different classes
26 according to source type, and that correlations between personal and ambient PM will be higher
27 when nonambient sources of PM are removed from the personal PM concentration. Mage (1998)
28 conducted analysis using the PTEAM data and showed that the average person in PTEAM
29 (Riverside, CA in the fall) was exposed to >75% of ambient PM2 5 and >64% of ambient PM10.
30 Mage et al. (1999) use an algorithm to fill in missing data and outliers to analyzed data sets and
April 2002 5-31 DRAFT-DO NOT QUOTE OR CITE
-------
>
TABLE 5-6. PAPERS REPORTING REANALYSES OF PARTICIPATE MATTER EXPOSURE STUDIES
Reference
Study Cited
Objectives/Hypotheses
Findings
to
o
to
75% of ambient PM25
and >64% of ambient PM10 measured by the community
monitor.
-------
TABLE 5-6 (cont'd). PAPERS REPORTING REANALYSES OF PARTICIPATE MATTER EXPOSURE STUDIES
to
o
o
to
Reference
Study Cited
Objectives/Hypotheses
Findings
Monn (2001)
Multiple Literature Review
To make an objective review of literature published since
1996 as an implicit update to Hie 1996 USEPA PM AQCD.
Emphasis on European studies.
Rotko et al. (2000a)
Rotko et al. (2000b)
Jantunen et al. (1998)
Carreretal. (1997)
Koistinen et al. (1999)
Rotko et al. (2000a),
Jantunen et al. (1998)
To make a comparison of exposure relationships between
the six EXPOLIS European cities (Athens, Basel, Grenoble,
Helsinki, Milan, Prague).
To determine sociodemographic influences of exposure in
Helsinki
"It is important to note that a personal measurement does
not a priori provide more valid data than a stationary
ambient measurement, i.e. a personal sample in a study
investigating effects from outdoor combustion particles is
often influenced by sources other than outdoor sources and
may thus confound the exposure-effect outcome."
"Despite some lack of correlation between personal (PM10)
and outdoor values, outdoor fine particle concentrations
were strongly associated with mortality and morbidity
indicating that outdoor sources (e.g. vehicular emissions)
emit the toxic entity (Dockerey et al., 1993; Schwartz et al.,
1996)."
Demographic Bias exists because women and more-
educated individuals are more likely to respond to survey.
Socioeconomic bias exists in low SES subjects less likely to
participate in diary keeping and exposure monitoring.
Weighting is required for inter-city comparisons.
Selection bias is not a problem for characterizing physical
factors influencing personal exposure
Distinct male vs female differences: Males had higher
exposures to PM2.5, related to ETS, and a larger variance
between sociodemographic groupings.
No sociodemographic differences existed in outdoor PM2.5
concentrations.
Lower occupational status contributed to greater P{M2.5
exposures than higher (professional) occupational status.
fe
H
6
o
o
H
O
O
H
W
O
o
-------
1 show that variation in daily personal exposures for subjects with similar activity patterns and no
2 ETS exposure are driven by variation in ambient PM concentrations.
3
4 5.4.3 Factors Influencing and Key Findings on Particulate Matter Exposures
5 5.4.3.1 Relationship of Personal/Microenvironmental Particulate Matter with Ambient
6 Particulate Matter
7 Understanding the relationship between ambient site measurements and personal exposure
8 to PM is important for several reasons. First, it allows us to determine when and for which PM
9 constituents it is appropriate to use ambient measurements as surrogates for exposure in
10 epidemiological studies. Second, it provides information that may improve surrogate exposure
11 measurements and, hence, increase the power of epidemiological studies. Finally, since
12 compliance with the NAAQS is based on ambient monitoring, it can be used to understand the
13 impact of regulation on exposures to PM and its constituents and, hence, can help link the impact
14 of regulations to health outcomes. Many of the studies, summarized above in Table 5-4, have
15 analyzed this relationship using measurements of personal PM exposures and ambient PM
16 concentrations. Of primary interest are the PM concentrations measured in ambient, indoor,
17 outdoor, and personal exposure samples, the statistical correlations between measurements, and
18 the attenuation and/or infiltration factors developed for personal exposure and indoor
19 microenvironments. Attenuation and infiltration factors are discussed in Section 5.3.2.3.1.
20 Information on correlation analysis are provided below.
21
22 5.4.3.1.1 Types of Correlations
23 The three types of correlation data that will be discussed in this section are longitudinal,
24 "pooled", and daily-average correlations. Longitudinal correlations are calculated when data
25 from a study includes measurements over multiple days for each subject (longitudinal study
26 design). Longitudinal correlations describe the temporal relationship between daily personal PM
27 exposure or microenvironment concentration and daily ambient PM concentration for each
28 individual subject. The longitudinal correlation coefficient, r, may differ for each subject, and an
29 analysis of the variability in r across subjects can be performed with this type of data. Typically,
30 the median r is reported along with the range across subjects in the study. There are two types of
31 cross-sectional correlations: pooled and daily average. Pooled correlations are calculated when
April 2002 5-34 DRAFT-DO NOT QUOTE OR CITE
-------
1 a study involves one or only a few measurements per subject and different subjects are studied on
2 subsequent days. Pooled correlations combine individual subject/individual day data for the
3 correlation calculation. Pooled correlations describe the relationship between daily personal PM
4 exposure and daily ambient PM concentration across all subjects in the study. For some studies,
5 the multiple days of measurements for each subject were assumed to be independent (after
6 autocorrelation and sensitivity analysis) and combined together in the correlation calculation
7 (Ebelt et al., 2000). Daily-average correlations are calculated by averaging exposure across
8 subjects for each day. Daily-average correlations then describe the relationship between the daily
9 average exposure and daily ambient PM concentration.
10 Pooled correlations have been simulated from longitudinal data by using a random -
11 sampling procedure to select a random day from each subject's measurements for use in the
12 correlation. This procedure was repeated many times, and statistics such as the mean and
13 standard deviation of the pooled correlation coefficient were reported (Janssen et al., 1997,
14 1998a, 1999c).
15 The type of correlation analysis can have a substantial effect on the resulting correlation
16 coefficient. Mage et al. (1999) mathematically demonstrated that very low correlations between
17 personal exposure and ambient concentrations could be obtained when people with very different
18 nonambient exposures are pooled, even though their individual longitudinal correlations are high.
19 The longitudinal studies conducted by Tamura et al. (1996a) and Janssen et al. (1997, 1998a,
20 1999c) determined that the longitudinal correlations between personal exposure and ambient PM
21 concentrations were higher than the correlations obtained from a pooled data set. Wallace
22 (2000a) reviewed a number of longitudinal studies and found that the median longitudinal
23 correlation coefficient was higher than the pooled correlation coefficient for the same data (see
24 Tables 1 and 2, Wallace, 2000a).
25 Mage et al. (1999) examined three longitudinal exposure data sets where several subjects
26 were measured each day. They showed that by averaging daily exposures across subjects, daily-
27 average correlations could be obtained. These were all higher than the median longitudinal
28 correlations. Williams et al. (2000a,b) and Evans et al. (2000) have also reported higher
29 correlation coefficients for daily-average correlations compared to longitudinal correlations. The
30 higher correlations found between daily-average personal exposures and ambient PM
31 concentrations, as opposed to lower correlations found between individual exposures and
April 2002 5-35 DRAFT-DO NOT QUOTE OR CITE
-------
1 ambient PM levels, recently have been attributed to the statistical process of averaging (Ott et al.,
2 2000). Personal exposures include contributions from nonambient as well as ambient PM
3 concentrations. When several subjects are measured on the same day the mean variability due to
4 variations in nonambient exposures are reduced due to averaging. Therefore, the correlation
5 between personal exposure and ambient concentrations increases as the number of subjects
6 measured daily increases. Ott et al. (2000), using the theory on which their Random Component
7 Superposition (RCS) model is based, predict expected correlations above 0.9 for the PTEAM
8 study and above 0.70 for the New Jersey study (Lioy et al., 1990) if 25 subjects had been
9 measured daily in each study.
10
11 5.4.3.1.2 Correlation Data from Personal Exposure Studies
12 Measurement data and correlation coefficients for the personal exposure studies described
13 in Section 5.4.2.1 are summarized in Table 5-7. All data are based on mass measurements. The
14 studies are grouped by the type of study design, longitudinal or pooled. For each study in
15 Table 5-7, summary statistics for the total personal PM exposure measurements are presented,
16 as well as statistics for residential indoor, residential outdoor, and ambient PM concentrations,
17 where available. The correlation coefficient (r) between total personal PM exposures and
18 ambient PM concentrations also are presented and classified as longitudinal or pooled
19 correlations. When reported,/^-values for the correlation coefficients are included. Correlation
20 coefficients between personal, indoor, outdoor, and ambient also are reported, when available.
21
22 5.4.3.1.3 Correlations Between Personal Exposures, Indoor, Outdoor, and Ambient
23 Measurements
24 Longitudinal and pooled correlations between personal exposure and ambient or outdoor
25 PM concentrations varied considerably between study and study subjects. Most studies report
26 longitudinal correlation coefficients that range from <0 to ~ 1, indicating that an individual's
27 activities and residence type may have a significant effect on total personal exposure to PM.
28 General population studies tend to show lower correlations because of the higher variation in the
29 levels of PM generating activities. In contrast, the absence of indoor sources for the populations
30 in several of the longitudinal studies resulted in high correlations between personal exposure and
31 ambient PM within subjects over time for these populations. But even for these studies,
April 2002 5-36 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-7. PERSONAL MONITORING STUDIES FOR PARTICIPATE MATTER: MEASURED CONCENTRATIONS
AND CORRELATION COEFFICIENTS
to
o
o
to
Measured Concentration Levels (,ug/m3)
Size
Fraction
Avg.
Time
Sample
Statistic Size1
Residential Residential
Personal Indoor Outdoor Ambient
Type3
Personal-Ambient 2 Other
Correlation Coefficients (r) Correlation Coefficients (r)
Value (Range)
Notes Type3 Value (Range)
Longitudinal Studies
Ebelt et al. (2
PM2.5
000) - Vancouver, BC
24 h
x±SD 106
Range
18.2 ±14.6 11.4 ±4.1
2-91 4 - 29
Median L
P
0.48 (-0.68-0.83)
0.15
n = 16 COPD
subjects
Evans et al. (2000) - Fresno, CA
i
OJ
O
?7
rti
H
6
o
2|
O
H
O
O
w
o
^
o
1 — I
H
W
PM2.5
PM25
Janssen et al.
PM10
Janssen et al.
PM10
Janssen et al.
PM2.5
PM2.5
PM25
24 h
24 h
(1997) -
24 h
x 24
Range
x 12
Range
Netherlands
x±SD 301
Range
13.3 9.7 20.5 21.7
1-24 4-17 4-52 6-37
11.1 8.0 10.1 8.6
7-16 4-12 5-20 4-16
105.2± 38.5 ±5.6
28.7 25 - 56
57-195
P
P
Median L
Median L
Median L
MeanP
MeanP
MeanP
0.414
0.844
0.63(0.1-0.9)
0.63
0.59
0.28 (0.12)5
0.45 (0.16)5
0.20 (0.19)5
Fresno-1 study Pp.; 0.814
PP-O 0.804
Fresno-2 study Pp.; 0.954
PP-O 0.804
n = 45 school
children
With nonsmoking
parents
With smoking
parents
All
With nonsmoking
parents
With smoking
parents
(1998a)- Netherlands
24 h
x±SD 262
Range
61.7 ±18.3 35.0 ±9.4 41.5 ±4.3
38-113 19-65 32-50
Median L
Median P
Median P
0.50 (-0.41-0.92)
0.50 (0.07-0. 83)5
0.34 (-0.09-
0.67)5
n = 37 adults Med. Lp.; 0.72 (-0.10-
No ETS exposure Med. L;.a 0.98)
All 0.73 (-0.88-
0.95)
(1999c)- Netherlands
24 h
24 h
24 h
x±SD 11
Range
x±SD 55
Range
x±SD 22
Range
28.3 ±11.3 17.1 ±2.8
19-60 14-22
24.4 ±4.9 17.1 ±2.6
19-33 15-22
37.0 ±17.4 17.1 ±3.7
21-60 14-21
Median L
Median P
Median L
Median P
0.86 (-0.1 1-0.99)
0.41 (-0.28-
0.93)5
0.92
0.825
n = 13 school
children
With nonsmoking
parents
With smoking
parents
-------
TABLE 5-7 (cont). PERSONAL MONITORING STUDIES FOR PM: MEASURED CONCENTRATIONS AND
CORRELATION COEFFICIENTS
to
o
o
to
Measured Concentration Levels (,ug/m3)
Size
Fraction
Janssen et al
PM25
Janssen et al
PM2.5
Avg.
Time
. (2000)
24 h
. (2000)
24 h
Statistic
- Netherlands
x±SD
Range
- Finland
x±SD
Range
Sample
Size1
338
336
Residential Residential
Personal Indoor Outdoor Ambient
24.3 ±25.7 28.6 ±41.8 20.6 ± 4.0
9-134 9-239 13-31
10.8 ±4.4 11.0 ±4.0 12.6 ±2.0
4-33 3-27 10-18
Personal-Ambient 2 Other
Correlation Coefficients (r) Correlation Coefficients (r)
Type3 Value (Range)
Median L 0.79 (-0.41-0.98)
Median L 0.85
Median L 0.76 (-0.12-0.97)
Notes Type3
n = 36 elderly Med. Lp.;
w/CV disease Med. L;..,
No ETS exposures
n = 46 elderly Med. Lp.;
w/CV disease Med. L^
Value (Range)
0.91 (-0.28-
1.0)
0.84(-0.00-
0.98)
0.89(0.14-1.0)
0.70 (-0.15-
0.94)
Linn et al. (1999) - Los Angeles
i
OJ
oo
PM25
PM10
24 h
24 h
Rojas-Bracho et al. (
O
(*
^
H
6
o
0
H
O
Cj
O
w
L-LJ
o
^
o
HH
H
W
PM25
PM10
PM10.2.5
Sarnat et al.
PM2.5
PM10
12 h
12 h
12 h
(2000) -
24 h
24 h
x±SD
Range
x±SD
Range
2000) - Boston
x±SD
Range
x±SD
Range
x±SD
Range
Baltimore
x±SD
x±SD
x±SD
x±SD
60
59
224
225
222
37
36
37
36
23.8 ±15.1 23.5 ±15.3 24.8 ± 14.5
4-65 4-92 4-63
34.8 ±14.8 32.6 ±15.6 39.8 ±18.3 33 ± 15
5-85 9-105 7-97 9-??
21.6 ±13.6 17.5±14.1 14.2 ±11.2
1-128 2-73 1-57
37.2 ±22.8 31.9 ±25.2 22.2 ±18.7
9-211 2-329 3-76
15.6 ±14.6 14.5 ±9.2 8.1 ±6.8
-11-103 -3-255 -2-64
26.7 ±13.7 25.2 ±11.5
18.5 ±11.2 5.6 ±49.0
33.9 ±11.7 34.0 ±12.8
28.0 ±16.5 7.5 ±73.2
P 0.266
P 0.226
Median L 0.61(0.10-0.93)'
Median L 0.35 (0.0-0.72)6
Median L 0.30 (0.0-0.97)6
Median L 0.76 (-0.21-
Median L 0.95)7
P 0.25 (-0.38-
P 0.81)7
P 0.898
P 0.758
0.508
0.448
Median L 0.64 (0.08-0. 86)7
Median L 0.53 (-0.79-
0.89)7
p
O
P™
PO-,
n = 17 adults Med. Lp.;
Med. L;.0
Med. Lp.;
Med. L;.0
Med. Lp.;
Med. L;.0
n= 15 adults;
summer
n= 15 adults;
winter
High ventilation;
summer
Med. ventilation;
summer
Low ventilation;
summer
WINTER
SUMMER
WINTER
0.266
0.476
0.326
0.666
0.876
0.746
0.716
0.506
0.426
0.206
-------
VO
H
6
o
o
H
O
O
HH
H
W
Personal-Ambient 2 Other
Measured Concentration Levels (,ug/m3) Correlation Coefficients (r) Correlation Coefficients (r)
Size Avg. Sample
Fraction Time Statistic Size1
Sarnat et al. (2000) - Baltimore (cont'd)
PM10.25 24 h x±SD 37
x±SD 36
Tamura et al. (1996a) - Tokyo
PM10 48 h
Williams et al. (2000a,b) - Baltimore
PM25 24 h x 23
Range
PM10 24 h x 28
Range
PM10.25 24 h x 26
Range
Residential Residential
Personal Indoor Outdoor Ambient Type3 Value (Range) Notes Type3
7.2 ±4.0 8.4 ±2.3 Median L 0.11 (-0.60- SUMMER
9.6 ±7.9 -1.3 ±24.2 Median L 0.64)7 WINTER
0.32 (-0.48-
0.68)7
P 0.83 n = 7 elderly adults
13.0 9.4 22.0 22.0 Median L 0.80(0.38-0.98)' n = 21 elderly Pp.;
7-25 4-19 7-52 8-59 P 0.894 adults Pp.0
P;-o
PO-!
11.0 30.0 29.9 P;.0
4-23 13-66 13 - 74 PiHl
1.0 8.0 8.0 P;.0
-3-5 -2-16 1-15 PiHl
PO-,
Value (Range)
0.904
0.954
0.94"
0.874
0.964
0.824
0.81"
0.944
0.184
0.084
0.45"
Pooled Studies
Bahadori (1998) - Nashville
PM25 12 h x±SD 30
Range
PM10 12 h x±SD 30
Range
Pellizzari et al. (1999) - Toronto
PM2.5 3 d x 922
PM10 3d x 141
21.7 ±10. 5 15.5 ±6.6 23.4 ±6.8 P 0.09 n=10COPD Pp.;
10-67 5-40 3-61 subjects; daytime P;.0
33.0 ±16.9 21.6 ±10.7 32.5 ± 8.1 P -0.08 n=10COPD Pp.;
5-88 9-77 7-76 subjects; daytime P;.0
28.4 21.1 15.1 P 0.23 n=178;nfor Pp.;
indoor, outdoor P;.0
lower than
personal
67.9 29.8 24.3 No correlations
reported
0.72
0.31
0.43
0.06
0.79
0.33
-------
>
to
o
o
to
TABLE 5-7 (cont). PERSONAL MONITORING STUDIES FOR PM: MEASURED CONCENTRATIONS AND
CORRELATION COEFFICIENTS
H
6
o
o
H
O
o
HH
H
W
Size Avg.
Fraction Time Statistic
Measured Concentration Levels (i/g/m3)
Sample Residential Residential
Size1 Personal Indoor Outdoor Ambient
Personal-Ambient 2 Other
Correlation Coefficients (r) Correlation Coefficients (r)
Type3 Value (Range) Notes Type3
Value (Range)
Oglesby et al. (2000) - EXPOLIS Basel
PM25 48 h x±SD
44 23.7 ±17.1 19.0 ±11.7
20 17.5 ±13.0 17.7 ±7.1
P 0.07 All
P 0.21 No ETS exposure
Santos-Burgoa et al. (1998) - Mexico City
PM10 24 h x±SD
Tamura et al. (1996b) - Osaka
PM2 48 h
PM10 48 h
66 97 ±44 99 ± 50
P 0.26 Pp.;
P 0.74
P 0.67
0.47
0.23
Pellizzari et al., 2001 - Indianapolis
PM25 72-h Median
250 23 18 18 18
P 0.102 Between the Pvs outdoor
Logarithms of
concentrations P vs Indoor
0.138
0.923
Brauer et al.,2000 - Banska BystricaSO4
PM25 24-h Mean
PM10
PM10 Summer 122 79 35
PM10 Winter 120 66 45
PM2 5 Summer 88 55 22
PM25 Winter 69 53 32
SO,, Winter 6.5 4.6 5.7
Abbreviations used: h = Hour
Avg. = Averaging (time) i-a = Indoor-ambient correlation
Cone. = Concentration i-o = Indoor-outdoor correlation
CV = Cardiovascular L = Longitudinal correlation
d = Day Med. = Median
ETS = Environmental tobacco smoke o-a = Outdoor-ambient correlation
P PM10 r2<0.17 Multivar. With P indoor
nicotine PM2.5
P
Personal SO4
vs Amb. SO4
P = Pooled correlation
p-i = Personal-indoor correlation
p-o = Personal-outdoor correlation
SD = Standard deviation
Stat. = Statistic
x = Mean
r2=0.15
r2=0.23
Notes:
1. Sample size is for personal concentrations; indoor, outdoor and ambient sample sizes may differ.
2. Correlation coefficient is for personal-residential outdoor if no ambient concentration data reported.
3. See text for description of types of correlations.
4. Daily-averaged correlation (values for individual subjects averaged for each day).
5. Pooled correlations estimated using a Monte Carlo sampling procedure, n = 1000. If mean P is shown, then SD given; if median P is shown, then range is given.
6. Obtained from a regression equation; r=^/(R2).
6. Spearman rank correlations.
7. Calculated, r = v(R2), from R2 from a mixed model regression.
-------
1 correlations varied by individual, depending on their activities and the microenvironments that
2 they occupied.
3
4 Probability Studies
5 In the Toronto study (Pellizzari et al., 1999), pooled correlations were derived for personal,
6 indoor, outdoor, and fixed site ambient measurements. This study was conducted in Toronto on
7 a probability sample of 732 participants who represented the general population, 16 years and
8 older. The study included between 185 and 203 monitoring periods with usable PM data for
9 personal, residential indoor, and outdoor measurements. For PM10, measurements, the mean
10 concentrations were 67.9 //g/m3 for personal, 29.8 //g/m3 for indoor air, and 24.3 //g/m3 for
11 outdoor air samples. For PM2 5, the mean concentrations were 28.4 //g/m3 for personal,
12 21.1 //g/m3 for indoor air, and 15.1 //g/m3 for outdoor air samples. A low but significant
13 correlation (r = 0.23, p < 0.01) was reported between personal exposure and ambient
14 measurements. The correlations between indoor concentrations and the various outdoor
15 measurements of PM25 ranged from 0.21 to 0.33. The highest correlations were for outdoor
16 measurements at the residences with the ambient measurements made at the roof site (0.88) and
17 the other fixed site (0.82). Pellizzari et al. (1999) state that much of the difference among the
18 data for personal/indoor/outdoor PM
19
20 ... can be attributed to tobacco smoking, since all variables reflecting smoking . . . were found to be
21 highly correlated with the personal (and indoor) particulate matter levels, relative to other variables that
22 were measured . . . none of the outdoor concentration data types (residential or otherwise) can
23 adequately predict personal exposures to particulate matter, (p. 729)
24
25 Using a Random Component Superposition Statistical Model, Ott et al. (2000) calculated
26 an attenuation factor of 0.6144 for personal exposure for PM10. The mean nonambient exposure
27 component for PM10 was estimated as 52.62 //g/m3 with a standard deviation of 84.82 //g/m3.
28 Although the data were available for PM2 5, similar calculations were not made.
29 PM10 data from the PTEAM study were analyzed using the same approach (Ott et al.,
30 2000). For PTEAM, an attenuation factor of 0.5546 was calculated for personal exposure.
31 Infiltration factors were calculated for each residence with an average of 0.5594 and a standard
April 2002 5-41 DRAFT-DO NOT QUOTE OR CITE
-------
1 deviation of 0.1476. Values ranged from a minimum of 0.19 to a maximum of 0.87 showing the
2 substantial variability that can be seen between homes depending upon the housing
3 characteristics and operation of the HVAC system. The mean nonambient exposure component
4 for PM10 was estimated as 59.23 //g/m3 with a standard deviation of 45.85 //g/m3.
5 Santos-Burgoa et al. (1998) describe a 1992 study of personal exposures and indoor
6 concentrations to a randomly sampled population near Mexico City. The sample of 66 monitored
7 subjects included children, students, office and industrial workers, and housewives. None of the
8 people monitored were more than 65 years old. The mean 24-h personal exposure and indoor
9 concentrations were 97 ± 44 (SD) and 99 ± 50 //g/m3, respectively, with an rPereonai/Ambient = 0.26
10 (p = 0.099). Other correlations of interest were rPereonaWndoor = 0.47 (p = 0.002) and rIndoor/Ambient =
11 0.23 (p = 0.158). A strong statistical association was found between personal exposure and
12 socioeconomic class (p = 0.047) and a composite index of indoor sources at the home
13 (p = 0.039).
14 Correlation analysis for personal exposure has not yet been reported for EXPOLIS. Some
15 preliminary results (Jantunen et al., 2000) show that, in Basel and Helsinki, a single ambient
16 monitoring station was sufficient to characterize the ambient PM25 concentration in each city.
17 Using microenvironmental concentration data collected while the subjects were at home, at work,
18 and outdoors, they calculated the sum of the time-weighted-averages of these data and found the
19 results closely match the personal PM25 exposure data collected by the monitors carried by most
20 of the subjects, with a few subjects (mostly smokers) being noticeable exceptions.
21
22 Longitudinal Studies
23 A number of longitudinal studies using a purposeful sampling design have been conducted
24 and reported in the literature since 1996. A number of these studies (Janssen et al., 1998a,
25 1999b, 2000; Williams et al., 2000b; Evans et al., 2000) support the previous work by Janssen
26 et al. (1995) and Tamura et al. (1996a) and demonstrate that, for individuals with little exposure
27 to nonambient sources of PM, correlations between total PM exposure and ambient PM
28 measurements are high. Other studies (Ebelt et al., 2000; Sarnat et al., 2000) show strong
29 correlations for the SO4"2 component of PM2 5 but poorer correlations for PM2 5 mass. Still other
30 studies show only weak correlations (Rojas-Bracho et al., 2000; Linn et al., 1999; Bahadori et al.,
31 2001). Even when strong longitudinal correlations are demonstrated for individuals in a study,
April 2002 5-42 DRAFT-DO NOT QUOTE OR CITE
-------
1 the variety of living conditions may lead to variations in attenuating factors or the fraction of
2 ambient PM contributing to personal exposure. Groups with similar living conditions, especially
3 if measurements are conducted during one season, may have similar a and, therefore, very high
4 correlations between personal exposure and ambient concentrations, even for pooled correlations.
5 However, when studies contain subjects with homes of very different ventilation characteristics
6 or cover more than one season, variations in a can be high across subjects, thus, showing poor
7 pooled correlations even in the absence of indoor sources.
8 Elderly Subjects. Janssen et al. (2000) continued their longitudinal studies with
9 measurements of personal, indoor, and outdoor concentrations of PM25 for elderly subjects with
10 doctor-diagnosed angina pectoris or coronary heart disease. Studies were conducted in
11 Amsterdam and Helsinki, Finland, in the winter and spring of 1998 and 1999. In the Amsterdam
12 study, with 338 to 417 observations, the mean PM2 5 concentrations were 24.3, 28.6, and 20.6
13 Mg/m3 for personal, indoor, and outdoor samples, respectively. If the measurements with ETS in
14 the home were excluded, the mean indoor concentration dropped to 16 //g/m3, which was lower
15 than outdoors. In the Helsinki study, the mean PM2 5 concentrations were 10.8 //g/m3 for
16 personal, 11.0 //g/m3 for indoor air, and 12.6 //g/m3 outdoor air samples. The authors note that
17 for this group of subjects, personal exposure, indoor concentrations, and ambient concentrations
18 of PM25 were highly correlated within subjects over time. Median Pearson's correlation
19 coefficients between personal exposure and outdoor concentrations were 0.79 in Amsterdam and
20 0.76 in Helsinki. The median Pearson's r for the indoor/outdoor relationship was 0.85 for the
21 Amsterdam study, excluding homes with ETS. The correlation for indoors versus outdoors was
22 0.70 for all homes.
23 Results from the correlation analysis can be used to estimate infiltration factors and
24 penetration factors for these two groups of subjects. In Amsterdam, the attenuation factor was
25 0.43 and the infiltration factor was 0.47. Very similar results were seen in Helsinki for the
26 attenuation factor (0.45) and the infiltration factor (0.51).
27 A series of PM personal monitoring studies involving elderly subjects was conducted in
28 Baltimore County, MD, and Fresno, CA. The first study was a 17-day pilot (January-February
29 1997) to investigate daily personal and indoor PMX 5 concentrations, and outdoor PM25 and
30 PM2 5_10 concentrations experienced by nonsmoking elderly residents of a retirement community
31 located near Baltimore (Liao et al., 1999; Williams et al., 2000c). The 26 residents were aged
April 2002 5-43 DRAFT-DO NOT QUOTE OR CITE
-------
1 65 to 89 (mean = 81), and 69% of them reported a medical condition, such as hypertension or
2 coronary heart disease. In addition, they were quite sedentary; less than 5 h day"1, on average,
3 was spent on ambulatory activities. Because most of the residents ate meals in a communal
4 dining area, the average daily cooking time in the individual apartments was only 0.5 h (range 0
5 to4.5h). About 96% of the residents' time was spent indoors (Williams et al., 2000c). Personal
6 monitoring, conducted for five subjects, yielded longitudinal correlation coefficients between
7 ambient concentrations and personal exposure ranging from 0.00 to 0.90.
8 The Baltimore main study and the Fresno study were conducted using similar monitoring
9 techniques and study design. Concentrations measured in these studies are summarized in
10 Table 5-8. For PM2 5, personal exposure and indoor air concentrations are similar for all three
11 studies even though outdoor air concentrations for Fresno in the winter are only half of those
12 measured for Fresno in the Winter and Baltimore in the summer. This result is presumably due
13 to high penetration efficiencies in the spring in Fresno when the weather was warm and
14 participants kept the windows and doors of their homes open. These data also show that even
15 when correlations are high, the use of an ambient monitor as a surrogate for exposure in
16 epidemiological studies can bias the strength of the health effect found, due to differing exposure
17 levels.
18
19
TABLE 5-8. MEAN CONCENTRATION FOR PM MASS REPORTED FOR THE
BALTIMORE (Williams et al., 2000a,b,c) AND THE FRESNO
(Evans et al., 2000) STUDIES
PM2 5 Concentration (/^g/m3) PM10 Concentration (/^g/m3)
Study
Baltimore
Fresno -Winter
Fresno-Spring
Personal
13.0 ±4.2
13. 3 ±5.9
11.1±2.8
Indoors
10.5 ±4.9
9.7 ±5.0
8.0 ±1.8
Outdoors Personal
22.0 ±12.0 —
20.5 ±13.4 —
10.1 ±3.2 37.3
Indoors
13. 5 ±6.3
15.1±4.1
16.7 ±3.1
Outdoors
30.0 ±13.7
28.2 ± 15.9
28.7 ±6.6
1 Calculated Correlation Coefficients are summarized in Tables 5-9 and 5-10. In Table 5-9,
2 results for Baltimore show excellent daily average correlations for both PM2 5 and PM10. These
April 2002 5-44 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-9. DAILY- AVERAGE CORRELATION COEFFICIENTS REPORTED
FOR THE BALTIMORE (WILLIAMS ET AL., 2000a,b,c) AND THE FRESNO
(EVANS ET AL., 2000) STUDIES
Study
Baltimore
Fresno -Winter
Fresno-Spring
Ambient/Outdoor
0.92
0.48
0.53
PM15 r2
Personal/Ambient
0.80
(0.14-0.80)a
—
0.70
Personal/Indoors
0.98
(0.20-0.99)a
—
0.77
PM,,,!-2
Ambient/Outdoor
0.89
0.48
0.61
aRange for individual participants.
TABLE 5-10. REGRESSION ANALYSIS REPORTED FOR INDOOR/OUTDOOR
RELATIONSHIPS FOR THE BALTIMORE (WILLIAMS ET AL., 2000a,b,c) AND
THE FRESNO (EVANS ET AL., 2000) STUDIES
Daily Average Individual
Study
Baltimore
Fresno -Winter
Fresno-Spring
r2
0.92
0.86
0.56
slope
0.39
nr
nr
Intercept
(/-tg/m3) r2 slope
1.5 0.73 ±0.16 0.43 ±0.15
nr 0.55 ±0.25 0.25 ±0.17
nr 0.39 ±0.21 0.49 ±0.38
Intercept
Cug/m3)
0.9 ±2.6
4.4 ±3.2
3.0 ±3.7
1 results represent primarily the behavior of fine particle regional sulfate for a group of participants
2 who have few indoor or personal sources. However even for this group, there was a wide range
3 of individual correlation coefficients. The Fresno data, on the other hand, shows much poorer
4 daily average correlations. Of special note are the poorer correlation for the ambient to outdoor
5 residential monitor. This could be due to the higher concentrations of nitrate in the samples. In
6 addition, the residential site may have be influenced by highway traffic.
7 The correlation analysis in Table 5-10 shows correlation coefficients as well as the slope
8 (infiltration factor) and the intercept (indoor concentration due to nonambient sources) for the
9 Baltimore and Fresno studies. These data show strongest correlations for Baltimore, with very
April 2002
5-45
DRAFT-DO NOT QUOTE OR CITE
-------
1 low indoor concentrations from nonambient sources. Correlations for Fresno are not as strong,
2 with higher concentrations from nonambient sources. The infiltration factors for Baltimore and
3 Fresno-Spring time are very similar at approximately 0.5. The infiltration factors for Fresno-
4 Winter are considerably lower.
5 Subjects with COPT). Linn et al. (1999) describe a 4-day longitudinal assessment of
6 personal PM25 and PM10 exposures (on alternate days) in 30 COPD subjects aged 56 to 83;
7 concurrent indoor and outdoor monitoring were conducted at their residences. This study
8 occurred in the summer and autumn of 1996 in the Los Angeles area. PM10 data from the nearest
9 fixed-site monitoring station to each residence also was obtained. Pooled correlations for
10 personal exposure to outdoor measurements were 0.26 and 0.22 for PM25 and PM10, respectively.
11 Correlations of day-to-day changes in PM2 5 and PM10 measured outside the homes and correlated
12 with concurrent PM10 measurements at the nearest ambient monitoring location gave R2 values of
13 0.22 and 0.44, respectively. Correlations of day to day changes in PM mass measured indoors
14 correlated with outdoor measurements at the homes gave R2 values of 0.27 and 0.19 for PM10 and
15 PM25, respectively.
16 Personal, indoor, and outdoor PM2 5 PM10 and PM2 5_10 correlations were reported by
17 Rojas-Bracho et al. (2000) for a study conducted in Boston, MA, on 18 individuals with COPD.
18 Both the mean and median personal exposure concentrations were higher than the indoor
19 concentrations, which were higher than outdoor concentrations for all three PM measurement
20 parameters. Geometric mean indoor/outdoor ratios were 1.4 ± 1.9 for PM10, 1.3 ± 1.8 for PM25,
21 and 1.5 ± 2.7 for PM2 5_10. Median longitudinal R2s between personal exposure and ambient PM
22 measurements were 0.12 for PM10, 0.37 for PM2 5 and 0.07 for PM2 5_10. The relationship between
23 the indoor and outdoor concentrations was strongest for PM2 5, with a median R2 of 0.55 and
24 11 homes having significant R2 values. For PM10 the median R2 value was 0.25, with significant
25 values for eight homes. Only five homes had significant indoor/outdoor associations for PM2 5_10,
26 with an insignificant median R2 value of 0.04. The poor correlations for PM10_2 5 are a result of
27 poorer penetration efficiencies, higher decay rates, and spatial inhomogeneities.
28 Bahadori et al. (2001) report a pilot study of the PM exposure of 10 nonrandomly chosen
29 chronic obstructive pulmonary disease (COPD) patients in Nashville, TN, during the summer of
30 1995. Each subject alternately carried a personal PM2 5 or PM10 monitor for a 12-h daytime
31 period (8 a.m. to 8 p.m.) for 6 consecutive days. These same pollutants were monitored
April 2002 5-46 DRAFT-DO NOT QUOTE OR CITE
-------
1 simultaneously indoors and outdoors at their homes. All of the homes were air-conditioned and
2 had low air exchange rates (mean = 0.57 h"1), which may have contributed to the finding that
3 mean indoor PM2 5 was 66% of the mean ambient PM2 5. This can be contrasted with the
4 PTEAM study in Riverside, CA, where no air conditioners were in use and the mean indoor
5 PM25 was 98% of the mean ambient PM25 (Clayton et al., 1993). Data sets were pooled for
6 correlation analysis. Resulting pooled correlations between personal and outdoor concentrations
7 were r = 0.09 for PM25 and r = -0.08 for PM10.
8
9 5.4.3.1.4 Sulfate as a Surrogate for Personal Exposure to Ambient Particulate Matter
10 A study, conducted in Vancouver, involving sixteen COPD patients aged 54 to 86, reported
11 low median longitudinal (r = 0.48) and pooled (r = 0.15) correlation coefficients between
12 personal exposures and ambient concentrations of PM25 (Ebelt et al., 2000). However, the mean
13 correlation between personal exposure to sulfate and ambient concentrations of sulfate was much
14 higher (r = 0.96). Since typically there are minimal indoor sources of sulfate, the relationship
15 between ambient concentrations and personal exposures to sulfate would not be weakened by
16 variability in an indoor-generated sulfate component, as for example in the case for PM2 5 for
17 which there are many primary indoor sources as well as some secondary indoor sources.
18 Correlations of ambient concentrations vs. personal exposures for PM2 5 and sulfate are compared
19 in Figure 5-1.
20 Another study, conducted in Baltimore, MD, involved 15 nonsmoking adult subjects
21 (>64 years old) who were monitored for 12 days during summer 1998 and winter 1999 (Sarnat
22 et al., 2000). All subjects (nonrandom selection) were retired, physically healthy, and lived in
23 nonsmoking private residences. Each residence, except one, was equipped with central
24 air-conditioning; however, not all residences used air-conditioning throughout the summer. The
25 average age of the subjects was 75 years (±6.8 years). Sarnat et al. (2000) reported higher
26 longitudinal and pooled correlations for PM2 5 during summer than winter. Similar to Ebelt et al.
27 (2000), Sarnat et al. (2000) reported stronger associations between personal exposure to SO42"
28 and ambient concentrations of SO42" than for total personal PM2 5 exposure and ambient PM2 5
29 concentrations. The ranges of correlations are shown in Figure 5-1 along with similar data from
30 Ebelt et al. (2000).
31
April 2002 5-47 DRAFT-DO NOT QUOTE OR CITE
-------
1.00 -
0.75 -
0 0.50 -
"c
0
'o 0.25 -
it
0
o
o o.oo -
c
O
| -0.25-
o
° -0.50 -
-0.75 -
1 nn
Ebeltetal., 2000
Pearson's "r"
e
2.5 37
T
Sulfate
Percentile
90th Percentile
75th Percentile
Median
25th Percentile
10th Percentile
Sarnatetal., 2000
Spearman's "r"
PM25
T
Y
I
r
—
1
raT
ll
T
Sulfate
PM2 5 Sulfate
PM2 5 Sulfate
Figure 5-1. Comparison of correlation coefficients for longitudinal analyses of personal
exposure versus ambient concentrations for individual subjects for VM2S and
sulfate.
1 The higher correlation coefficients and the narrower range of the correlation coefficient for
2 sulfate suggest that removing indoor-generated and personal activity PM from total personal PM
3 would result in a higher correlation with ambient concentrations. If there are indeed no indoor
4 sources, a personal exposure measurement for sulfate gives the ambient exposure of sulfate; the
5 ratio of personal sulfate to ambient sulfate gives the attenuation coefficient on an individual,
6 daily basis; and the attenuation coefficient times the ambient PM2 5 concentration gives the
7 individual, daily values of ambient PM2 5 exposures (Wilson et al., 2000).
8 This technique applies only to the non-volatile components of fine PM, as measured by
9 PM2 5. It requires that the sulfate concentration be large enough so that it can be measured with
10 reasonable accuracy. It does not require that sulfate be correlated with PM2 5 or the non-sulfate
11 components of PM25 since the sulfate data is used to estimate the attenuation coefficient, not
April 2002
5-48
DRAFT-DO NOT QUOTE OR CITE
-------
1 PM2 5. The technique does require that there be minimal indoor sources of sulfate, as indicated
2 by a near-zero intercept for the regression, and that the size distribution of PM25 and sulfate be
3 similar.
4 Sarnat et al. (2001) subsequently extended the Baltimore study to include 20 older adults,
5 21 children, and 15 individuals with COPD for a total of 56 subjects. In both studies (Sarnat
6 et al., 2000, 2001), they used their personal and ambient sulfate data to estimate the ambient
7 PM2 5 exposure. They used this information in mixed model analysis (mixed models account for
8 differences among individual subjects) but did not report correlations between ambient PM25
9 exposure and ambient PM2 5 concentrations.
10 However, Sarnat et al. (2001) did report slopes from the mixed model analyses. The
11 t-statistic for the slope of ambient exposure versus ambient concentration as compared to total
12 personal exposure versus ambient concentration increased from 9.96 to 11.12 (total exposure vs.
13 ambient concentration) for the summer period and 4.36 to 19.88 (ambient exposure vs. ambient
14 concentration) for the winter period.
15 The study conducted by Sarnat et al. (2000) also illustrates the importance of ventilation on
16 personal exposure to PM. During the summer, subjects recorded the ventilation status of every
17 visited indoor location (e.g., windows open, air-conditioning use). As a surrogate for the
18 air-exchange rate, personal exposures were classified by the fraction of time the windows were
19 open while a subject was in an indoor environment (Fv). Sarnat et al. (2000) report regression
20 analyses for personal exposure on ambient concentration for total PM2 5 and for sulfate for each
21 of the three ventilation conditions. Personal exposure to sulfate may be taken as a surrogate for
22 personal exposure to ambient accumulation-mode PM in the absence of indoor sulfate sources.
23 Figure 5-2 shows a comparison of the regressions and indicates how the use of a sulfate tracer as
24 a surrogate for PM of ambient origin improves the correlation coefficient. The improvement is
25 especially pronounced for the lowest ventilation conditions. For the lowest ventilation condition,
26 R2 improves from 0.25 to 0.72.
27
28 5.4.3.1.5 Personal Exposure to Ambient andNonambient Particulate Matter
29 The utility of treating personal exposure to ambient PM, Eag, and personal exposure to
30 nonambient PM, Enonag, as separate and distinct components of total personal exposure to PM, Et,
31 was pointed out by Wilson and Suh (1997). The PTEAM study measured, in addition to indoor,
April 2002 5-49 DRAFT-DO NOT QUOTE OR CITE
-------
60
E 50-|
D)
CD" 40-
t/i
a so H
20-
Well Ventilated Indoor Environment
35
R2 = 0.80
30
25
20
15
10
5-
R2 = 0.88
60
50-
of 40-
U)
I 3°
LU
CD
C
o
U)
20-
10-
Moderately Vented Indoor Environment
35-
R2 = 0.57
X
• X
X
30-
25-
20-
15-
10-
5-
R2 = 0.73
60
E 50-
IL
CD" 40-
01
g. 30-|
.
.2
20-
o
S2
CD 1n I
Q_ 10 -
Poorly Ventilated Indoor Environment
35
R2 = 0.25
30-
25-
20-
15-
10-
5-
R2 = 0.72
0 10
20 30 40
PM2.5
50 60 0 5 10
15 20 25
S04
30 40
Ambient Concentration (|jg/m3)
Figure 5-2. Personal exposure versus ambient concentrations for PM2 5 and sulfate. (Slope
estimated from mixed models).
Source: Sarnat et al. (2000).
April 2002
5-50
DRAFT-DO NOT QUOTE OR CITE
-------
1 outdoor, and personal PM, the air exchange rate for each home and collected information on the
2 time spent in various indoor and outdoor ^e. This information is available for 147, 12-h daytime
3 periods. With this information, it is possible to estimate the daytime Eag and Enonag as described
4 in Section 5.3.2.3.1. Various examples of this information have been reported (Mage et al.,
5 1999; Wilson et al., 2000). Graphs showing the relationships between ambient concentration and
6 the various components of personal exposure (Et, Eag, and Enonag) are shown in Figure 5-3. The
7 correlation coefficient for the pooled data set improves from r = 0.377 for Et versus Ca
8 (Figure 5-3a) to r = 0.856 for Eag versus Ca (Figure 5-3b) because of the removal of the Enonag ,
9 which, as shown in Figure 5-3c, is highly variable and independent of Ca. The correlation
10 between Eag and Ca is less than 1 because of the day-to-day variation in ait. The regression
11 analysis with Et total PM gives (X = 0.711 and N = 81.6 //g/m3. The regression analysis with Eag
12 gives oT = 0.625. The regression with Enonag gives N = 79.2 //g/m3. The finite intercept in the
13 regression with Eag must be attributed to bias or error in some of the measurements. No studies,
14 other than PTEAM, have provided the quantity of data on Et, Ca, C;, and a required to conduct
15 an analysis comparable to that shown in Figure 5-3. It should be noted that the PTEAM study
16 was conducted in southern California in the fall, when house were open and air exchange rates
17 were high and relatively uniform. These are best case conditions for showing high correlations
18 between ambient site measurements and personal correlations.
19 The RCS model introduced by Ott et al.(2000) presents a modeling framework to determine
20 the contribution of ambient PM10 and indoor-generated PM10 on personal exposures in large
21 urban metropolitan areas. The model has been tested using personal, indoor and outdoor PM10
22 data from three urban areas (Riverside, CA; Toronto; and Phillipsburg, NJ). Results suggest that
23 it is possible to separate the ambient and nonambient PM contributions to personal exposures on
24 a community-wide basis. However, as discussed in the paper, the authors make some
25 assumptions that require individual consideration in each-city specific application of the model
26 for exposure or health effects investigations. Primarily, housing factors, air-conditioning,
27 seasonal differences, and complexities in time-activity profiles specific to the cohort being
28 studied have to be taken into account prior to adopting the model to a given situation. Finally,
29 this and other available exposure-based analyses presented here does not yet predict the relative
30 contribution of indoor and outdoor PM to particle mass burden to the lung as a function of
April 2002 5-51 DRAFT-DO NOT QUOTE OR CITE
-------
250
100
si 75~
O
13 .
(/) DJ
O ro
Q.LJJ
UJ5 50H
« .2
fc£
°-| 25H
0
200
r =0.856
R"= 0.733
A= 1.16+0.625C
«=147
5150 H
s ^
2 I
|jio°H
UJ?
53 E
Q_;A * *
* ** V • •
i i i
1
50 100 150 200
Ambient Concentration, |jg/m3
250
Figure 5-3. Regression analyses of aspects of daytime personal exposure to PM10 estimated
using data from the PTEAM study, (a) Total personal exposure to PM, Et,
regressed on ambient concentration, Ca. (b) Personal exposure to ambient PM,
Eag regressed on Ca. (c) Personal exposure to nonambient PM, Enonag regressed
onCa.
Source: Data taken from Clayton et al. (1993).
April 2002
5-52
DRAFT-DO NOT QUOTE OR CITE
-------
1 human activities and different microenvironmental sources and concentrations of PM and its
2 copollutants.
3
4 5.4.3.2 Factors That Affect Relationship between Personal Exposure and Ambient PM
5 A number of factors will affect the relationship between personal exposure and PM
6 measured at ambient-site community monitors. Spatial variability in outdoor microenvironments
7 and penetration into indoor microenvironments will influence the relationship for ambient-
8 generated PM, air-exchange rates, and decay rates in indoor microenvironments will influence
9 the relationship for both ambient-generated and total PM, whereas personal activities will
10 influence the relationship for total PM but not ambient-generated PM. Information on these
11 effects is presented in detail in the following section.
12
13 5.4.3.2.1 Spatial Variability and Correlations Over Time
14 Chapter 3 (Section 3.2.3) presents information on the spatial variability of PM mass and
15 chemical components at fixed-site ambient monitors; for purposes of this chapter, this spatial
16 variability is called an "ambient gradient". The data presented in Section 3.2.3 indicate that
17 ambient gradients of PM and its constituents exist in urban areas to a greater or lesser degree.
18 This gradient, and any that may exist between a fixed-site monitor and the outdoor jue near where
19 people live, work, and play, obviously affects the exposure. The purpose of this section is to
20 review the available data on ambient monitor-to-outdoor microenvironmental concentration
21 gradients, or relationships, that have been measured by researchers since 1996. These analyses
22 below are, in general, consistent with the previous studies covered in the 1996 PM AQCD. A
23 few outdoor-to-outdoor monitoring studies also are included to highlight relationships among
24 important jue categories. To assess spatial variability or gradients, the spatial correlations in the
25 data are usually analyzed. However, it should be noted that high temporal correlation between
26 two monitoring locations does not imply low spatial variability or low ambient gradients. High
27 temporal correlation between two sites indicates that changes in concentrations at one site can be
28 estimated from data at another site. Results presented below are consistent with previous studies
29 assessed in the 1996 PM AQCD.
30 Oglesby et al. (2000), in a paper on the EXPOLIS-EAS study, conclude that very little
31 spatial variability exists in Basel, Switzerland, between PM levels measured at fixed site
April 2002 5-53 DRAFT-DO NOT QUOTE OR CITE
-------
1 monitors and the participant's outdoor jue. The authors report a high correlation between home
2 outdoor PM25 levels (48-h measurements beginning and ending at 8:00 a.m.) and the
3 corresponding 24-h average PM4 (time-weighted values calculated from midnight to midnight)
4 measured at a fixed monitoring station (n = 38, rsp = 0.96, p < 0.001). They considered each
5 home outdoor monitor as a temporary fixed monitor and concluded that "the PM2 5 level
6 measured at home outdoors . . . represents the fine particle level prevailing in the city of Basel
7 during the 48-h measuring period . . . ."
8 In a study conducted in Helsinki, Finland, Buzorius et al. (1999) conclude that a single
9 monitor may be used to adequately describe the temporal variations in concentration across the
10 metropolitan area. Particle size distributions were measured using a differential mobility particle
11 sizer (DMPS; Wintlmayer) coupled with a condensation particle counter (CPC TSI 3010, 3022)
12 at four locations including the official air monitoring station, which represented a "background"
13 site. The monitoring period varied between 2 weeks and 6 mo for the sites and data were
14 reported for 10-min and 1-, 8-, and 24-h averages. As expected, temporal variation decreased as
15 the averaging time increased. The authors report that particle number concentration varied in
16 magnitude with local traffic intensity. Linear correlation coefficients computed for all possible
17 site-pairs and averaging times showed that the correlation coefficient improved with increasing
18 averaging time. Using wind speed and direction vectors, lagged correlations were calculated and
19 were generally higher than the "raw" data correlations. Weekday correlations were higher than
20 weekend correlations as "traffic provides relatively uniform spatial distribution of particulate
21 matter" (p. 565). The authors conclude that, even for time periods of 10 min and 1 h, sampling at
22 one station can describe temporal variations across relatively large areas of the city with a
23 correlation coefficient >0.7.
24 Dubowsky et al. (1999) point out that, although the variation of PM25 mass concentration
25 across a community may be small, there may be significant spatial variations of specific
26 components of the total mass on a local scale. An example is given of a study of concentrations
27 of polycyclic aromatic hydrocarbons (PAH) at three indoor locations in a community:
28 (1) an urban and (2) a semi-urban site separated by 1.6 km, and (3) a suburban site located further
29 away. The authors found the geometric mean PAH concentrations at these three locations varied
30 respectively as 31:19:8 ng/m3, and suggest that the local variations in traffic density were
31 responsible for this gradient. Note that these concentrations are 1,000 times lower than the total
April 2002 5-54 DRAFT-DO NOT QUOTE OR CITE
-------
1 PM mass concentration, so that such a small gradient would not be detectable for total PM2 5
2 mass measurements on the order of 25 //g m"3.
3 The THEES study reported by Waldman et al. (1991) measured indoor, outdoor, and
4 personal BaP levels and found that the outdoor BaP was the same at all outdoor sites across the
5 three sampling periods. This study showed the seasonal differences vs. BaP levels and exposures
6 due to indoor and outdoor sources and individual activities.
7 Leaderer et al. (1999) monitored 24-h PM10, PM2 5, and sulfates during the summers of
8 1995 and 1996 at a regional site in Vinton, VA (6 km from Roanoke, VA). One similar 24-h
9 measurement was made outdoors at residences in the surrounding area, at distances ranging from
10 1 km to >175 km from the Vinton site, at an average separation distance of 96 km. The authors
11 reported significant correlations for PM2 5 and sulfates between the residential outdoor values and
12 those measured at Vinton on the same day. In addition, the mean values of the regional site and
13 residential site PM25 and sulfates showed no significant differences in spite of the large distance
14 separations and mountainous terrain intervening in most directions. However, for the
15 concentrations of PM25.10, estimated as PM10-PM2 5, no significant correlation among these sites
16 was found (n = 30, r = -0.20).
17 Lillquist et al. (1998) found no significant gradient in PM10 concentrations in Salt Lake
18 City, UT, when levels were low, but a gradient existed when levels were high. PM10
19 concentrations were measured outdoor at three hospitals using a Minivol 4.01 sampler
20 (Airmetrics, Inc.) operating at 5 L min"1 and at the Utah Department of Air Quality (DAQ)
21 ambient monitoring station located between 3 and 13 km from the hospitals for a period of about
22 5 mo.
23 Pope et al. (1999) monitored ambient PM10 concentrations in Provo, UT (Utah Valley),
24 during the same time frame the following year and reported nearly identical concentrations at
25 three sites separated by 4 to 12 km. Pearson correlation coefficients for the data were between
26 0.92 and 0.96. The greater degree of variability in the Salt Lake City PM10 data relative to the
27 Provo data may be related to the higher incidence of wind-blown crustal material in Salt Lake
28 City. Pope et al. (1999) reported that increased health effects in the Utah Valley were associated
29 with stagnation and thermal inversions trapping anthropogenically derived PM10, whereas, no
30 increases in health effects were observed when PM10 levels were increased during events of wind
31 blown crustal material.
April 2002 5-55 DRAFT-DO NOT QUOTE OR CITE
-------
1 Vakeva et al. (1999) found significant vertical gradients in submicron particles existed in
2 an urban street canyon of Lahti, Finland. Particle number concentrations were measured using a
3 TSI screen diffusion battery and a condensation particle counter at 1.5 and 25 m above the street
4 at rooftop level. The authors found a fivefold decrease in concentration between the two
5 sampling heights and attributed the vertical gradient to dilution and dispersion of pollutants
6 emitted at street level.
7 White (1998) suggests that the higher random measurement error for the coarse PM
8 fraction compared to the error for the fine PM fraction may be responsible for a major portion of
9 the apparent greater spatial variability of coarse ambient PM concentration compared to fine
10 ambient PM concentration in a community (e.g., Burton et al., 1996; Leaderer et al., 1999).
11 When PM2 5 and PM10 are collected independently, and the coarse fraction is obtained by
12 difference (PM25_10 = PM10-PM25), then the expected variance in the coarse fraction is influenced
13 by the variances of the PM10 and PM25 measurements. When a dichotomous sampler collects
14 PM25 and PM25.10 on two separate filters, the coarse fraction also is expected to have a larger
15 error than the fine fraction. There is a possible error caused by loss of mass below the cut-point
16 size and a gain of mass above the cut-point size that is created by the asymmetry of the product
17 of the penetration times PM concentration about the cut-point size. Because a dichotomous PM
18 sampler collects coarse mass using an upper and lower cut-point, it is expected to have a larger
19 variance than for fine mass collected using only one cut-point.
20 Wilson and Suh (1997) conclude that PM25 and PM10 concentrations are correlated more
21 highly across Philadelphia than are PM2 5_10 concentrations. Ambient monitoring data from 1992
22 to 1993 was reviewed for PM2 5, PM2 5_10, and PM10, as well as for PM25 and PM25.10 dichotomous
23 data for 212 site-years of information contained in the AIRS database (U.S. Environmental
24 Protection Agency, 2000). The authors also observed that PM10 frequently was correlated more
25 highly with PM2 5 than with PM2 5_10. The authors note that PM2 5 constitutes a large fraction of
26 PM10, and that this is the likely reason for the strong agreement between PM2 5 and PM10. Similar
27 observations were made by Keywood et al. (1999) in six Australian cities. The authors reported
28 that PM10 was more highly correlated with PM2 5 than with coarse PM (PM2 5_10), suggesting that
29 "variability in PM10 is dominated by variability in PM2 5."
30 Lippmann et al. (2000) examined the site-to-site temporal correlations in Philadelphia
31 (1981 to 1994) and found the ranking of median site-to-site correlation was O3 (0.83), PM10
April 2002 5-56 DRAFT-DO NOT QUOTE OR CITE
-------
1 (0.78), TSP (0.71), NO2 (0.70), CO (0.50), and SO2 (0.49). The authors explain that O3 and a
2 fraction of TSP and PM10 (e.g., sulfate) are secondary pollutants that would tend to be distributed
3 spatially more uniformly within the city than primary pollutants such as CO and SO2, which are
4 more likely to be influenced by local emission sources. Lippman et al. (2000) conclude: "Thus,
5 spatial uniformity of pollutants may be due to area-wide sources, or to transport (e.g., advection)
6 of fairly stable pollutants into the urban area from upwind sources. Relative spatial uniformity of
7 pollutants would therefore vary from city to city or region to region."
8
9 5.4.3.2.2 Physical Factors Affecting Indoor Microenvironmental Particulate Matter
10 Concentrations
11 Several physical factors affect ambient particle concentrations in the indoor
12 microenvironment, including air exchange, penetration, and particle deposition. Combined,
13 these factors are critical variables that describe ambient particle dynamics in the indoor jue and,
14 to a large degree, significantly affect an individual's personal exposure to ambient-generated
15 particles while indoors. The relationship between ambient outdoor particles and ambient
16 particles that have infiltrated indoors is given by:
17
18 Cai/Cao=Pa/(a + k\ (5-12)
19
20 where Cai and Cao are the concentration of ambient indoor and outdoor particles, respectively;
21 /"is the penetration factor; a is the air exchange rate; and k is the particle deposition rate (as
22 discussed in Section 5.3.2.3.1, use of this model assumes equilibrium conditions and assumes
23 that all variables remain constant). Particle penetration is a dimensionless quantity that describes
24 the fraction of ambient particles that effectively penetrates the building shell. "Air exchange" is
25 a term used to describe the rate at which the indoor air in a building or residence is replaced by
26 outdoor air. The dominant processes governing particle penetration are air exchange and
27 deposition of particles as they traverse through cracks and crevices and other routes of entry into
28 the building. Although air-exchange rates have been measured in numerous studies, very few
29 field data existed prior to 1996 to determine size-dependent penetration factors and particle
30 deposition rates. All three parameters (P, a, and K) may vary substantially depending on building
31 type, region of the country, and season. In the past several years, researchers have made
April 2002 5-57 DRAFT-DO NOT QUOTE OR CITE
-------
1 significant advancements in understanding the relationship between particle size and penetration
2 factors and particle deposition rates. This section will highlight the studies that have been
3 conducted to better understand physical factors affecting indoor particle dynamics.
4
5 Air-Exchange Rates
6 The air-exchange rate, a, in a residence varies depending on a variety of factors, including
7 geographical location, age of the building, the extent to which window and doors are open, and
8 season. Murray and Burmaster (1995) used measured values of a from households throughout
9 the United States to describe empirical distributions and to estimate univariate parametric
10 probability distributions of air-exchange rates. Figure 5-4 shows the results classified by season
11 and region. In general, a is highest in the warmest region and increases from the coldest to the
12 warmest region during all seasons. Air-exchange rates also are quite variable within and between
13 seasons, as well as between regions (Figure 5-4). Data from the warmest region in summer
14 should be viewed cautiously as many of the measurements were made in Southern California in
15 July, when windows were more likely to be open than in other areas of the country where
16 air-conditioning is used. Use of air-conditioning generally results in lowering air-exchange rates.
17 In a separate analyses of these data, Koontz and Rector (1995) suggested that a conservative
18 estimate for air exchange in residential settings would be 0.18 h"1 (10th percentile) and a typical
19 air exchange would be 0.45 h"1 (50th percentile).
20 These data provide reasonable experimental evidence that a varies by season in locations
21 with distinct seasons. As a result, infiltration of ambient particles may be more efficient during
22 warmer seasons when windows are likely to be opened more frequently and air-exchange rates
23 are higher. This suggests that the fraction of ambient particles present in the indoor (j,e would be
24 greater during warmer seasons than colder seasons. For example, in a study conducted in
25 Boston, MA, participants living in non-air-conditioned homes kept the windows closed except
26 during the summer (Long et al., 2000a). This resulted in higher and more variable air-exchange
27 rates in summer than during any other season (Figure 5-5). During nighttime periods, when
28 indoor sources are negligible, the indoor/outdoor concentration ratio or infiltration factor may be
29 used to determine the relative contribution of ambient particles in the indoor microenvironment.
30 Particle data collected during this study (Figure 5-6) shows the indoor/outdoor concentration
31 ratios by particle size. Data show that, for these nine homes in Boston, the fraction of ambient
April 2002 5-58 DRAFT-DO NOT QUOTE OR CITE
-------
03
o:
(D
O)
C
03
O
X
LU
3 -
2 -
1 -
0
Coldest Region
Colder Region
Warmer Region
Warmest Region
I
I
Winter Spring Summer
Season
Fall
Figure 5-4. Air-exchange rates measured in homes throughout the United States. Climatic
regions are based on heating-degree days: Coldest region > 7000, Colder
region = 5500 to 6999, Warmer region = 2500 to 4999, and Warmest region
< 2500 heating-degree days.
Based on data from Murray and Burmaster (1995).
1
2
3
4
5
particles penetrating indoors is higher during summer when air exchange rates were higher than
fall (Long et al., 2000b).
Particle Deposition Rates and Penetration Factors
Physical factors affecting indoor particle concentrations, including particle deposition rates,
k, and penetration factors, P, are possibly the most uncertain and variable quantities. Although k
can be modeled with some success, direct measurements are difficult and results often vary from
study to study. Particle deposition rates vary considerably depending on particle size because of
the viscous drag of air on the particles hindering their movement to varying degrees. The nature
April 2002
5-59
DRAFT-DO NOT QUOTE OR CITE
-------
&
.0)
CD
O '
X
LU
? 1 -
0 -
Fall
95%
90%
Median
Winter
Spring
Season
Summer
Figure 5-5. Box plots of hourly air-exchange rates stratified by season in Boston, MA,
during 1998.
Source: Long et al. (2000a).
O
ra
1.1
1.0 -
0.9 -
0.8 -
0.7 -
0.6 -
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
0.0
8
9
8
o
p
o
o
o
0.1
Summer Fall
§
o
•>- IO CM
O T d
% ° £
odd
co ^r iq
pop
CN| CO -4
odd
CO
CN
o
CD
Particle Diameter (|jm)
Figure 5-6. Geometric mean infiltration factor (indoor/outdoor ratio) for hourly nighttime,
nonsource data for two seasons. Box plots of air exchange rates are shown as
inserts for each plot. (Boston, 1998)
Source: Longetal. (2000b).
April 2002
5-60
DRAFT-DO NOT QUOTE OR CITE
-------
1 and composition of particles also affect deposition rates. Surface properties of particles, such as
2 their electrostatic properties, can have a significant influence on deposition rates. In addition,
3 thermophoresis can also affect k, but probably to a lesser degree in the indoor (j,e because
4 temperatures generally vary over a small range. Combined, these effects can produce order of
5 magnitude variations in k between particles of different size and, in the case of electrophoresis
6 and thermophoresis, particles of the same size.
7 Particle penetration efficiency into indoor microenvironments depends on particle size and
8 air exchange rates. Penetration varies with particle size because of the size-dependent deposition
9 of particles caused by impact!on, interception, and diffusion of particles onto surfaces as they
10 traverse through cracks and crevices. Penetration also is affected by air exchange rates. When
11 air exchange rates are high, P approaches unity because the majority of ambient particles have
12 less interaction with the building shell. In contrast, when air exchange rates are low, P is
13 governed by particle deposition as particles travel through cracks and crevices.
14 Significant advancements have been made in the past few years to better characterize
15 particle deposition rates and penetration factors. Several new studies, including two in which
16 semi-continuous measurements of size distributions were measured indoors and outdoors, have
17 produced new information on these quantities, which are key to understanding the contributions
18 of ambient PM to indoor PM concentrations (Equation 5-10).
19 Studies involving semi-continuous measurements of indoor and outdoor particle size
20 distributions have been used to estimate k and P as a function of particle size (Vette et al., 2001;
21 Long et al., 2000b; Abt et al., 2000b). These studies each demonstrated that the indoor/outdoor
22 concentration ratios (Cai/Cao in Equation 5-12) were highest for accumulation mode particles and
23 lowest for ultrafme and coarse-mode particles. Various approaches were used to estimate size-
24 specific values for k and P. Vette et al. (2001) and Abt et al. (2000b) estimated k by measuring
25 the decay of particles at times when indoor levels were significantly elevated. Vette et al. (2001)
26 estimated P using measured values of k and indoor/outdoor particle measurements during
27 nonsource nighttime periods. Long et al. (2000b) used a physical-statistical model, based on
28 Equation 5-10, to estimate k and P during nonsource nighttime periods. The results for k
29 reported by Long et al. (2000b) and Abt et al. (2000b) are compared with other studies in
30 Figure 5-7. Although not shown in Figure 5-7, the results for k obtained by Vette et al. (2001)
31 were similar to the values of & reported by Abt et al. (2000b) for particle sizes up to 1 //m.
April 2002 5-61 DRAFT-DO NOT QUOTE OR CITE
-------
-
10-
-
\ '.
32
m :
o
I*
r
I
I
error bar
includes 0
1
5
i
10
Particle size (|jm)
aDecay rates represent Summary Estimates from the four houses examined.
bDecay rates are based on sujfate and are presented as <2.5 urn.
Estimates were computed using a surface-to-volume ratio of 2 rrr1 (Koutrakis ef a/., 1992).
cData represent PM25
dPartide sizes are the midpoint of the ranges examined.
"Decay rates presented are estimates of k for nightly average data from all nine study homes.
fDecay rates are theoretically modeled deposition values for smooth indoor surfaces and homogeneous and isotropically turbulent air flow.
Presented curves assume typical room dimensions (3 m x 4 m x 5 m) and a friction velocity of 1.0 cm/s.
Figure 5-7. Comparison of deposition rates from this study with literature values (adapted
from Abt et al., 2000b). Error bars represent standard deviations for same-
study estimates.
Source: Adapted from Long et al. (2000b).
1
2
3
4
5
Results for P by Long et al. (2000b) show that penetration was highest for accumulation-mode
particles and decreased substantially for coarse-mode particles (Figure 5-8). The results for
P reported by Vette et al. (2001) show similar trends, but are lower than those reported by Long
et al. (2000b). This likely is because of lower air-exchange rates in the Fresno, CA, residence
(a ~ 0.5 h4; Vette et al., 2001) than the Boston, MA, residences (a > 1 h'1; Long et al., 2000b).
These data for P and k illustrate the role that the building shell may provide in increasing the
April 2002
5-62
DRAFT-DO NOT QUOTE OR CITE
-------
o
g>
'o
£
LJJ
(0
•5
c
0)
a.
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
o o
(SI CO
O O
o o
•st-
p
o
p
in
co
O
(N
O
o
O
-4
o
in CD
-4 in
Size Interval (|jm)
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
TO
o:
U)
o
Q.
0)
Q
Figure 5-8. Penetration efficiencies and deposition rates from models of nightly average
data. Error bars represent standard errors. (Boston, 1998, winter and
summer)
Source: Longetal. (2000b).
1
2
3
4
5
9
10
concentration of particles because of indoor sources and reducing the concentration of indoor
particles from ambient sources, especially for homes with low air-exchange rates.
Compositional Differences Between Indoor-Generated and Ambient-Generated
Particulate Matter
Wilson et al. (2000) discuss the differences in composition between particles from indoor
and outdoor sources. They note that, because of the difficulty in separating indoor PM into
ambient and nonambient PM, there is little direct experimental information on the composition
differences between the two. Although experimental data are limited, Wilson et al. (2000)
suggest the following.
April 2002
5-63
DRAFT-DO NOT QUOTE OR CITE
-------
1 Photochemistry is significantly reduced indoors; therefore, most secondary sulfate [H2S04,
2 NH4HS04, and (NH4)2S04] and nitrate (NH4N03) found indoors come from ambient sources.
3 Primary organic emissions from incomplete combustion may be similar, regardless of the source.
4 However, atmospheric reactions of polyaromatic hydrocarbons and other organic compounds
5 produce highly oxygenated and nitrated products, so these species are also of ambient origin.
6 Gasoline, diesel fuel, and vehicle lubricating oil all contain naturally present metals or metal
7 additives. Coal and heavy fuel oil also contain more metals and nonmetals, such as selenium and
8 arsenic, than do materials such as wood or kerosene burned inside homes. Environmental
9 tobacco smoke (ETS), however, with its many toxic components, is primarily an indoor-generated
10 pollutant.
11 Particles generated indoors may have different chemical and physical properties than those
12 generated by anthropogenic ambient sources. Siegmann et al. (1999) have demonstrated that
13 elemental carbon in soot particles generated indoors have different properties than in those
14 generated outdoors by automotive or diesel engines. In the United States, combustion-product
15 PM in the ambient/outdoor air generally is produced by burning fossil fuels (e.g., coal, gasoline,
16 fuel oil) and wood, whereas combustion-product PM from indoor sources is produced by
17 biomass burning (e.g., tobacco, wood, foods, etc.). However, some indoor sources of PM (such
18 as cigarette smoking, meat cooking, and coal burning) occur both indoors and outdoors and may
19 constitute an identifiable portion of measured ambient PM (Cha et al., 1996; Kleeman and Cass,
20 1998).
21
22 Indoor Air Chemistry
23 Gas- and aerosol-phase chemical reactions in the indoor microenvironment are responsible
24 for secondary particle formation and modification of existing particles. This process could be
25 complex and may influence the interpretation of exposures to indoor generated particles in
26 instances when particles are generated by outdoor gases reacting with gases indoors to produce
27 fresh particles. For example, homogeneous gas phase reactions involving ozone and terpenes
28 (specifically d-limonene, a-terpinene, and a-pinene) have been identified as an important source
29 of submicron particles (Weschler and Shields, 1999). Terpenes are present in several commonly
30 available household cleaning products and d-limonene has been identified in more than 50% of
31 the buildings monitored in the BASE study (Hadwen et al., 1997). Long et al. (2000a) found that
32 when PineSol (primary ingredient is a-pinene) was used indoors, indoor PM2 5 mass
April 2002 5-64 DRAFT-DO NOT QUOTE OR CITE
-------
1 concentrations increased by 3 to 32 //g m"3 (indoor ozone concentrations unknown, but ambient
2 ozone concentrations were 44 to 48 ppb). Similarly, a 10-fold increase in number counts of 0.1
3 to 0.2 //m particles was observed in an experimental office containing supplemented d-limonene
4 and normally encountered indoor ozone concentrations (< 5 to 45 ppb), resulting in an average
5 increase in particle mass concentration of 2.5 to 5.5 //g m"3 (Weschler and Shields, 1999). Ozone
6 appears to be the limiting reagent as particle number concentration varied proportionally to ozone
7 concentrations (Weschler and Shields, 1999). Other studies showed similar findings (e.g., Jang
8 and Kamens, 1999; Wainman et al., 2000).
9
10 Indoor Sources of Particles
11 The major sources of indoor PM in nonsmoking residences and buildings include
12 suspension of PM from bulk material, cooking, cleaning, and the use of combustion devices,
13 such as stoves and kerosene heaters. Human and pet activities also lead to PM detritus
14 production (from tracked-in soil, fabrics, skin and hair, home furnishings, etc.), which is found
15 ubiquitously in house dust deposited on floors and other interior surfaces. House dust and lint
16 particles may be resuspended indoors by agitation (cleaning) and turbulence (HVAC systems,
17 human activities, etc.). Ambient particles that have infiltrated into the indoor ^e also may be
18 resuspended after deposition to indoor surfaces. Typically, resuspension of particles from any
19 source involves coarse-mode particles (>1 //m); particles of smaller diameter are not resuspended
20 efficiently. On the other hand, cooking produces both fine- and coarse-mode particles, whereas
21 combustion sources typically produce fine-mode particles.
22 Environmental tobacco smoke (ETS) is also a major indoor source of PM. It is, however,
23 beyond the scope of this chapter to review the extensive literature on ETS. A number of articles
24 provide source strength information for cigarette or cigar smoking (e.g., Daisey et al., 1998 and
25 Nelson et al., 1998).
26 A study conducted on two homes in the Boston metropolitan area (Abt et al., 2000a)
27 showed that indoor PM sources predominate when air exchange rates were <1 h"1, and outdoor
28 sources predominate when air exchange rates were >2 h"1. The authors attributed this to the fact
29 that when air-exchange rates were low (<1 h"1), particles released from indoor sources tend to
30 accumulate because particle deposition is the mechanism governing particle decay and not air
31 exchange. Particle deposition rates are generally <1 h"1, especially for accumulation-mode
April 2002 5-65 DRAFT-DO NOT QUOTE OR CITE
-------
1 particles. When air-exchange rates were higher (>2 h"1), infiltration of ambient aerosols and
2 exfiltration of indoor-generated aerosols occur more rapidly, reducing the impact of indoor
3 sources on indoor particle levels. The study also confirmed previous findings that the major
4 indoor sources of PM are cooking, cleaning, and human activity. They discuss the size
5 characteristics of these ubiquitous sources and report the following.
6
7 The size of the particles generated by these activities reflected their formation processes.
8 Combustion processes (oven cooking, toasting, and barbecuing) produced fine particles and
9 mechanical processes (sauteing, frying, cleaning, and movement of people) generated coarse
10 particles. These activities increased particle concentrations by many orders of magnitude higher
11 than outdoor levels and altered indoor size distributions. (Abt et al., 2000a; p. 43)
12
13 They also note that variability in indoor PM for all size fractions was greater than for outdoor
14 PM, especially for short averaging times (2 to 33 times higher).
15 In a separate study conducted in nine nonsmoking homes in the Boston area, Long et al.
16 (2000a) concluded that the predominant source of indoor fine particles was infiltration of outdoor
17 particles, and that cooking activities were the only other significant source of fine particles.
18 Coarse particles, however, had several indoor sources, such as cooking, cleaning, and various
19 indoor activities. This study also concluded that more than 50% of the particles (by volume)
20 generated during indoor events were ultrafme particles. Events that elevated indoor particle
21 levels were found to be brief, intermittent, and highly variable, thus requiring the use of
22 continuous instrumentation for their characterization. Table 5-11 provides information on the
23 mean volume mean diameter (VMD) for various types of indoor particle sources. The
24 differences in mean VMD confirm the clear separation of source types and suggest that there is
25 very little resuspension of accumulation-mode PM. In addition, measurements of organic and
26 elemental carbon indicated that organic carbon had significant indoor sources, whereas elemental
27 carbon was primarily of ambient origin.
28 Vette et al. (2001) found that resuspension was a significant indoor source of particles
29 >1 //m, whereas fine- and accumulation-mode particles were not affected by resuspension.
30 Figure 5-9 shows the diurnal variability in the indoor/outdoor aerosol concentration ratio from an
31 unoccupied residence in Fresno. The study was conducted in the absence of common indoor
April 2002 5-66 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-11. VOLUME MEAN DIAMETER (VMD) AND MAXIMUM PM25
CONCENTRATIONS OF INDOOR PARTICLE SOURCESab
Size Statistics
Particle Source
Cooking
Baking (Electric)
Baking (Gas)
Toasting
Broiling
Sauteing
Stir-Frying
Frying
Barbecuing
Cleaning
Dusting
Vacuuming
Cleaning with Pine Sol
General Activities
Walking Vigorously (w/Carpet)
Sampling w/Carpet
Sampling w/o Carpet
Burning Candles
N
8
24
23
4
13
o
J
20
2
11
10
5
15
52
26
7
Indoor Activity
Mean VMD
(Mm)
0.189f
0.107f
0.138f
0.1 14f
0.184f, 3.48g
0.135f
0.173f
0.159f
5.38g
3.86B
0.097f
3.96g
4.25g
4.28B
0.311f
Background3'6
Mean VMD
(urn)
0.221f
0.224f
0.222f
0.236f
0.223f, 2.93g
0.277f
0.223f
0.205f
3.53g
2.79g
0.238f
3.18g
2.63g
2.93g
0.224f
Maximum
Mean
14.8
101.2
54.9
29.3
65.6
37.2
40.5
14.8
22.6
6.5
11.0
12.0
8.0
4.8
28.0
PM25
Concentration°'d
SD
7.4
184.9
119.7
43.4
95.4
31.4
43.2
5.2
22.6
3.9
10.2
9.1
6.6
3.0
18.0
Notes:
aAll concentration data corrected for background particle levels.
Includes only individual particle events that were unique for a given time period and could be detected above
background particle levels.
°PM concentrations in ^g/m3.
dMaximum concentrations computed from 5-min data for each activity.
background data are for time periods immediately prior to the indoor event.
fSize statistics calculated for PV002.05 using SMPS data.
gSize statistics calculated for PV0 7_10 using APS data.
Source: Long et al. (2000a).
1
2
3
4
particle sources such as cooking and cleaning. The data in Figure 5-9 show the mean
indoor/outdoor concentration ratio for particles >1 //m increased dramatically during daytime
hours. This pattern was consistent with indoor human activity levels. In contrast, the mean
April 2002
5-67
DRAFT-DO NOT QUOTE OR CITE
-------
0.0
06
Time
Figure 5-9. Mean hourly indoor/outdoor particle concentration ratio from an unoccupied
residence in Fresno, CA, during spring 1999.
Source: Vetteetal. (2001).
indoor/outdoor concentration ratio for particles <1 //m (fine- and accumulation-mode particles)
remain fairly constant during both day and night.
4 5.4.3.2.3 Time/Activity Patterns
5 Total exposure to PM is the sum of various microenvironmental exposures that an
6 individual encounters during the day and will depend on the microenvironments occupied.
7 As discussed previously, PM exposure in each microenvironment is the sum of exposures from
8 ambient sources (Eag), indoor sources (Epig), and personal activities (Epact). Eag and Epig are
9 determined by the microenvironments in which an individual spends time; whereas Epact is
10 determined by the personal activities that he/she conducts while in those microenvironments.
11 As mentioned before, PM exposures and its components are variable across the population; and,
April 2002
5-68
DRAFT-DO NOT QUOTE OR CITE
-------
1 thus, each are distributions rather than point estimates. A thorough analyses of these
2 distributions would require a comprehensive sensitivity and uncertainty analysis.
3 Determining microenvironments and activities that contribute significantly to human
4 exposure begins with establishing human activity pattern information for the general population,
5 as well as subpopulations. Personal exposure and time activity pattern studies have shown that
6 different populations have varying time activity patterns and, accordingly, different personal PM
7 exposures. Both characteristics will vary greatly as a function of age, health status, ethnic group,
8 socioeconomic status, season, and region of the country. Collecting detailed time activity data
9 can be very burdensome on participants but is clearly valuable in assessing human exposure and
10 microenvironments. For modeling purposes, human activity data frequently come from general
11 databases that are discussed below.
12 The gathering of human activity information, often called "time-budget" data, started in the
13 1920s; however, their use for exposure assessment purposes only began to be emphasized in the
14 1980s. Many of the largest U.S. human activity databases have been consolidated by EPA's
15 National Exposure Research Laboratory's (NERL) into one comprehensive database containing
16 over 22,000 person-days of 24-h activity known as the Consolidated Human Activity Database,
17 or CHAD (Glen et al., 1997; McCurdy et al., 2000). The information in CHAD is accessible for
18 constructing population cohorts of people with diverse characteristics that are useful for analysis
19 and modeling (McCurdy, 2000). See Table 5-2 for a summary listing of human activity studies
20 in CHAD. Most of the databases in CHAD are available elsewhere, including the National
21 Human Activity Pattern Survey (NHAPS), California's Air Resources Board (CARB), and the
22 University of Michigan's Institute for Survey Research data sets.
23 Although CHAD provides a very valuable resource for time and location data, there is less
24 information on PM-generating personal activities. In addition, very few of the time-activity
25 studies have collected longitudinal data within a season or over multiple seasons. Such
26 longitudinal data are important in understanding potential variability in activities and how they
27 impact correlations between PM exposure and ambient site measurements for both total PM and
28 PM of ambient origin.
29
30
April 2002 5-69 DRAFT-DO NOT QUOTE OR CITE
-------
1 5.4.3.3 Impact of Ambient Sources on Exposures to Particulate Matter
2 Different sources may generate ambient PM with different aerodynamic and chemical
3 characteristics, which may, in turn, result in different health responses. Thus, to fully understand
4 the relationship between PM exposure and health outcome, exposure from difference sources
5 should be identified and quantified. Source apportionment techniques provide a method for
6 determining personal exposure to PM from specific sources. Daily contributions from sources
7 that have no indoor component can be used as tracers to generate exposure estimates for ambient
8 PM of similar aerodynamic size or directly as exposure surrogates in epidemiologic analyses.
9 The most recent EPA PM Research Needs Document (U. S. Environmental Protection Agency,
10 1998) recommended use of source apportionment techniques to determine daily time-series of
11 source categories for use in community, time-series epidemiology.
12 A number of epidemiological studies (discussed more fully in Chapter 8) have evaluated
13 relationships between health outcomes and sources of particulate matter determined from
14 measurements at a community monitor. These studies suggest the importance of examining
15 sources and constituents of indoor, outdoor, and personal PM. For example, Ozkaynak and
16 Thurston (1987) evaluated the relationship between PM sources and mortality in 36 Standard
17 Metropolitan Statistical Areas (SMSAs). Particulate matter samples from EPA's Inhalable
18 Particle (IP) Network were analyzed for SO42" and NO3" by automated colorimetry, and elemental
19 composition was determined with X-ray fluorescence (XRF). Mass concentrations from five PM
20 source categories were determined from multiple regression of absolute factor scores on the mass
21 concentration: (1) resuspended soil, (2) auto exhaust, (3) oil combustion, (4) metals, and (5) coal
22 combustion.
23 In another study, Mar et al. (2000) applied factor analysis to evaluate the relationship
24 between PM composition (and gaseous pollutants) in Phoenix. In addition to daily averages of
25 PM2 5 elements from XRF analysis, they included in their analyses organic and elemental carbon
26 in PM2 5 and gaseous species emitted by combustion sources (CO, NO2, and SO2). They
27 identified five factors classified as (1) motor vehicles, (2) resuspended soil, (3) vegetative
28 burning, (4) local SO2, and (5) regional sulfate.
29 Also, Laden et al. (2000) applied specific rotation factor analysis to particulate matter
30 composition (XRF) data from six eastern cities (Ferris et al., 1979). Fine PM was regressed on
April 2002 5-70 DRAFT-DO NOT QUOTE OR CITE
-------
1 the recentered scores to determine the daily source contributions. Three main sources were
2 identified: (1) resuspended soil (Si), (2) motor vehicle (Pb), and (3) coal combustion (Se).
3 Source apportionment or receptor modeling has been applied to the personal exposure data
4 to understand the relationship between personal and ambient sources of particulate matter.
5 Application of source apportionment to ambient, indoor, and personal PM composition data is
6 especially useful in sorting out the effects of particle size and composition. If a sufficient
7 number of samples are analyzed with sufficient compositional detail, it is possible to use
8 statistical techniques to derive source category signatures, identify indoor and outdoor source
9 categories and estimate their contribution to indoor and personal PM.
10 Positive Matrix Factorization (PMF) has been applied to the PTEAM database by
11 Yakovleva et al. (1999). The authors utilize mass and XRF elemental composition data from
12 indoor and outdoor PM2 5 and personal, indoor, and outdoor PM10 samples. PMF is an advance
13 over ordinary factor analysis because it allows measurements below the quantifiable limit to be
14 used by weighting them by their uncertainty. This effectively increases the number of species
15 that can be used in the model. The factors used by the authors correspond to general source
16 categories of PM, such as outdoor soil, resuspended indoor soil, indoor soil, personal activities,
17 sea-salt, motor vehicles, nonferrous metal smelters, and secondary sulfates. PMF, by identifying
18 not only the various source factors but also apportioning them among the different monitor
19 locations (personal, indoor, and outdoor), was able to quantify an estimate of the contribution of
20 resuspended indoor dust to the personal cloud (15% from indoor soil and 30% from resuspended
21 indoor soil). Factor scores for these items then were used in a regression analysis to estimate
22 personal exposures (Yakovleva et al., 1999).
23 The most important contributors to PM10 personal exposure were indoor soil, resuspended
24 indoor soil, and personal activities; these accounted for approximately 60% of the mass
25 (Yakovleva et al., 1999). Collectively, they include personal cloud PM, smoking, cooking, and
26 vacuuming. For both PM2 5 and PM10, secondary sulfate and nonferrous metal operations
27 accounted for another 25% of PM mass. Motor vehicle exhausts, especially starting a vehicle
28 inside of an attached garage, accounted for another 10% of PM mass. The authors caution that
29 these results may not apply to other geographic areas, seasons of the year, or weather conditions.
30 Simultaneous measurement of personal (PM10) and outdoor measurements (PM25 and
31 PM10) were evaluated as a three-way problem with PMF, which allowed for differentiation of
April 2002 5-71 DRAFT-DO NOT QUOTE OR CITE
-------
1 source categories based on their variation in time and type of sample, as well as their variation in
2 composition. By use of this technique, it was possible to identify three sources of coarse-mode,
3 soil-type PM. One was associated with ambient soil, one with indoor soil dispersed throughout
4 the house, and one with soil resulting from the personal activity of the subject.
5 Two other source apportionment models have been applied to ambient measurement data
6 and can be used for the personal exposure studies. The effective variance weighted Chemical
7 Mass Balance (CMB) receptor model (Watson et al., 1984, 1990, 1991) solves a set of linear
8 equations that incorporate the uncertainty in the sample and source composition. CMB requires
9 the composition of each potential source of PM and the uncertainty for the sources and ambient
10 measurements. Source apportionment with CMB can be conducted on individual samples;
11 however, composition of each of the sources of PM must be known. An additional source
12 apportionment model, UNMIX (Henry et al., 1994) is a multivariate source apportionment
13 model. UNMIX is similar to PMF, but does not use explicitly the measurement uncertainties.
14 Because measurement uncertainties are not used, only species above the detection limit are
15 evaluated in the model. UNMIX provides the number of sources and source contributions and
16 requires a similar number of observations as PMF.
17 The Yakovleva et al. (1999) study demonstrates that source apportionment techniques also
18 could be very useful in determining parameters needed for exposure models and for determining
19 exposure to ambient-generated PM. Exposure information, similar to that obtained in the
20 PTEAM study, but including other PM components useful for definition of other source
21 categories (e.g., elemental [EC] and organic carbon [OC]; organic tracers for elemental carbon
22 from diesel vehicle exhaust, gasoline vehicle exhaust, and wood combustion; nitrate; Na; Mg and
23 other heavy metal tracers; and, also, gas-phase pollutants) would be useful as demonstrated in the
24 use of EC/OC and gas-phase pollutants by Mar et al. (2000).
25
26 5.4.3.4 Correlations of Particulate Matter with Other Pollutants
27 Several epidemiological studies have included the gaseous pollutants CO, NO2, SO2, and
28 O3 along with PM10 or PM25 in the analysis of the statistical association of health responses with
29 pollutants. In a recent study, the personal exposure to O3 and NO2 were determined, as well as
30 that to PM25 and PM25.10 for a cohort 15 elderly subjects in Baltimore, MD, although measured
31 personal exposures to O3, NO2, and SO2 were below their respective LOD for 70% of the
April 2002 5-72 DRAFT-DO NOT QUOTE OR CITE
-------
1 samples. Spearman correlations for 14 subjects in summer and 14 subjects in winter are given in
2 Table 5-12 for relationships between personal PM2 5 and ambient concentrations of PM2 5,
3 PM2 5_10, O3, and NO2. In contrast to ambient concentrations, neither personal exposure to total
4 PM2 5 nor PM2 5 ambient origin was correlated significantly with personal exposures to the
5 co-pollutants, PM2 5_10, nonambient PM2 5, O3, NO2, and SO2. Personal-ambient associations for
6 PM2 5_10, O3, NO2, and SO2 were similarly weak and insignificant. Based on these results, Sarnat
7 et al. (2000) conclude that the potential for confounding of PM25 by O3, NO2, or PM 10_2 5 appears
8 to be limited, because, despite significant correlations observed among ambient pollutant
9 concentrations, the correlations among personal exposures were low.
10 Sarnat et al. (2001) further evaluated the role of gaseous pollutants in particulate matter
11 epidemiology by extending the measurements taken on the earlier adult cohort of 20 individuals
12 in Baltimore by including additional PM and gaseous pollutant measurements that were
13 collected during the same 1998-1999 period from 15 individuals with chronic obstructive
14 pulmonary disease (COPD) and from 21 children. 24-h average personal exposures for PM25, O3,
15 SO2 and NO2, and corresponding ambient concentrations for PM2 5, O3, SO2, NO2 and CO for all
16 56 subjects were collected over 12 consecutive days. Results from correlation and regression
17 analysis of the personal and ambient data showed that personal PM2 5 and personal gaseous
18 pollutant exposures were generally not correlated. The analysis also showed that ambient PM25
19 concentrations had significant associations with personal PM25 exposures in both seasons. On the
20 other hand, ambient gaseous pollutant concentrations were not correlated with their
21 corresponding personal exposure concentrations. However, ambient gaseous concentrations were
22 found to be strongly associated with personal PM2 5 exposures, suggesting that ambient gaseous
23 concentrations for O3, NO2, SO2 are acting as surrogates, as opposed to confounders of PM25, in
24 the estimation of PM health effects based on multi pollutant models. This study did not measure
25 personal CO and also did not find a significant association between summertime ambient CO and
26 personal PM2 5 (a significant winter time association, however, was found). Personal EC and SO4
27 were also measured during the winter for the cohort of COPD patients only. The analysis of this
28 subset of the data showed that personal SO4 was significantly and negatively associated with
29 ambient O3 and SO2, and personal EC was significantly associated with ambient O3, NO2 and
30 CO. The authors interpret these findings as suggesting that O3 is primarily a surrogate for
31 secondary particle exposures, whereas ambient CO and NO2 are primarily surrogates for particles
April 2002 5-73 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-12. CORRELATIONS BETWEEN PERSONAL PM2 5 AND AMBIENT
POLLUTANT CONCENTRATIONS1
Personal PM2 5
vs. Ambient:
SUMMER Subject
SA1
SA2
SA5
SB1
SB2
SB3
SB4
SB5
SB6
SCI
SC2
SC3
SC4
SC5
WINTER WAI
WA2
WA4
WAS
WB1
WB2
WB3
WB4
WC1
WC2
WC3
WC4
WC5
WC6
Median Summer
Median Winter
PM25
0.55
0.85
0.89
0.65
-0.21
0.82
0.73
0.73
0.53
0.95
0.78
0.85
0.78
0.55
0.22
-0.38
-0.18
0.22
0.80
0.62
0.55
-0.12
0.74
0.79
0.28
0.19
0.81
0.01
0.76
0.25
03
0.15
0.31
0.18
0.40
-0.62
0.55
0.62
0.45
0.15
0.78
0.68
0.78
0.66
0.51
-0.18
-0.07
0.67
-0.43
-0.84
-0.32
-0.45
-0.01
-0.62
-0.88
-0.42
-0.84
-0.62
-0.03
0.48
-0.43
NO2
0.38
0.66
0.82
-0.15
0.81
-0.14
-0.34
-0.42
-0.38
0.66
0.36
0.73
0.59
0.32
-0.26
-0.36
-0.22
0.61
0.77
0.59
0.62
0.34
-0.15
0.17
0.03
0.50
0.08
0.65
0.37
0.26
PM2,.10
-0.12
0.57
0.64
0.38
0.15
-0.04
-0.12
0.23
0.12
0.65
0.51
0.68
0.70
0.43
-0.05
-0.70
-0.29
0.50
0.41
0.09
0.04
-0.10
0.44
0.77
0.57
0.45
0.81
0.37
0.41
0.39
Personal PM
of Ambient Origin vs
03
0.27
0.21
0.33
0.89
0.26
0.52
0.45
0.36
-0.03
0.83
0.66
0.69
0.50
0.34
-0.78
-0.15
-0.33
-0.72
-0.87
-0.76
-0.77
-0.80
-0.64
-0.87
-0.77
-0.72
-0.76
-0.75
0.41
-0.76
NO2
0.71
0.64
0.81
-0.74
0.08
-0.20
-0.29
-0.48
-0.57
0.63
0.65
0.71
0.50
0.33
-0.04
-0.15
0.20
-0.09
0.53
0.59
0.56
0.68
0.02
0.25
0.30
0.22
0.05
0.19
0.42
0.21
2.5
. Ambient:
PM2,.10
0.15
0.68
0.79
-0.03
0.33
0.00
-0.14
0.33
0.32
0.57
0.76
0.80
0.51
0.27
-0.24
0.02
0.00
0.40
0.66
0.59
0.60
0.48
0.69
0.71
-0.45
0.67
0.42
-0.45
0.33
0.45
'Correlations represent Spearman's rvalues; italicized values indicate significance at the a = 0.05 level.
Source: Sarnat et al. (2000).
April 2002
5-74
DRAFT-DO NOT QUOTE OR CITE
-------
1 from traffic. Sarnat et.(2001) caution that these findings were found in only one location and
2 various physical and personal factors, such as ventilation, time spent outdoors, and household
3 characteristics could affect the strength of the reported associations for certain individuals and
4 cohorts, even though the qualitative results found are unlikely to change.
5 A newly developed Roll-Around System (RAS) was used to evaluate the hourly
6 relationship between gaseous pollutants (CO, O3, NO2, SO2, and VOCs) and PM (Chang et al.,
7 2000). Exposures were characterized over a 15-day period for the summer and winter in
8 Baltimore, based on scripted activities to simulate activities performed by older adults (65+ years
9 of age). Spearman rank correlations were reported for PM2 5, O3, CO, and toluene for both the
10 summer and winter and the correlations are given for each microenvironment in Table 5-13:
11 indoor residence, indoor other, outdoor near roadway, outdoor away from road, and in vehicle.
12 No significant relationships (p < 0.05) were found between hourly PM2 5 and O3. Significant
13 relationships were found between hourly PM25 and CO: indoor residence, winter; indoor other,
14 summer and winter; and outdoor away from roadway, summer. Significant relationships also
15 were found between hourly PM2 5 and toluene: indoor residence, winter; indoor other, winter;
16 and in vehicle, winter. The significant relationships between CO and PM2 5 in the winter may be
17 caused by reduced air-exchange rates that could allow them to accumulate (Chang et al., 2000).
18 Although no significant correlation was found between in vehicle PM25 and CO, toluene, which
19 is a significant component of vehicle exhaust (Conner et al., 1995), was correlated significantly
20 to PM2 5 in the winter.
21 Carrer et al. (1998) present data on the correlations among personal and
22 microenvironmental PM10 exposures and concentrations and selected environmental chemicals
23 that were monitored simultaneously (using a method that was not described). These chemicals
24 were nitrogen oxides (NOX), carbon monoxide (CO), and total volatile organic compounds
25 (TVOC), benzene, toluene, xylene, and formaldehyde. The Kendall u correlation coefficient was
26 used; only results significant at p < 0.05 are mentioned here. Significant associations were found
27 only between the following pairs of substances (T shown in parentheses): personal PM10 (24 h)
28 andNOx (0.34), CO (0.34), TVOC (0.18), toluene (0.19), and xylene (0.26); office PM10 andNOx
29 (0.31); home PM10 and NOX (0.24), CO (0.24), toulene (0.17), and xylene (0.25). Surprisingly,
30 because most of the chemical substances are associated with motor vehicular emissions, there
April 2002 5-75 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-13. CORRELATIONS BETWEEN HOURLY PERSONAL PM2 5 AND
GASEOUS POLLUTANTS
Indoor
Residence
PM2.5vs.
Summer
Winter
PM2.5vs.
Summer
Winter
PM2.svs.
Summer
Winter
N
03
35
56
CO
41
59
Toluene
46
66
rs
0.29
0.05
0.25
0.43a
0.23
0.38a
Indoor Other
N
16
37
19
39
21
47
rs
-0.14
-0.06
0.59a
0.62a
-0.14
0.44a
Outdoor Near
Roadway
N
10
11
13
13
14
17
rs
0.05
-0.28
0.14
0.37
0.26
0.40
Outdoor Away
from road
N
12
7
12
8
14
8
rs
0.45
0.04
0.62
0.41
0.02
0.48
In Vehicle
N
37
34
46
37
48
42
rs
0.21
-0.10
0.23
0.10
0.12
0.43a
aCorrelations represent Spearman's rvalues; italicized values indicate significance at the a, = 0.05 level.
Source: Chang et al. (2000).
1
2
3
4
5
6
7
8
9
10
11
12
13
was no significant correlation between "commuting PM10" and any of the substances (Carrer
etal., 1998).
5.5 SUMMARY OF PARTICULATE MATTER CONSTITUENT DATA
5.5.1 Introduction
Atmospheric PM contains a number of chemical constituents that may be of significance
with respect to the human exposure and health effects. These constituents may be either
components of the ambient particles or bound to the surface of particles. They may be elements,
inorganic species, or organic compounds. A limited number of studies have collected data on
concentrations of elements, acidic aerosols, and polycyclic aromatic hydrocarbons (PAHs) in
ambient, personal, and microenvironmental PM samples. But, there have not been extensive
analyses of the constituents of PM in personal or microenvironmental samples. Data from
April 2002
5-76
DRAFT-DO NOT QUOTE OR CITE
-------
1 relevant studies are summarized in this section. The summary does not address bacteria,
2 bioaerosols, viruses, or fungi (e.g., Owen et al., 1992; Ren et al., 1999).
3
4 5.5.2 Monitoring Studies That Address Particulate Matter Constituents
5 A limited number of studies have measured the constituents of PM in personal or
6 microenvironmental samples. Relevant studies published in recent years are summarized in
7 Tables 5-11 and 5-12 for personal exposure measurements of PM and microenvironmental
8 samples, respectively. Studies that measured both personal and microenvironmental samples are
9 included in Table 5-11.
10 The largest database on personal, microenvironmental, and outdoor measurements of PM
11 elemental concentrations is the PTEAM study (Ozkaynak et al., 1996b). The results are
12 highlighted in the table and discussed below. The table shows that a number of studies have
13 measured aerosol acidity, sulfate, ammonia, and nitrate concentrations. Also, a number of
14 studies have measured PAHs, both indoors and outdoors. Other than the PAHs, there is little
15 data on organic constituents of PM.
16
17 5.5.3 Key Findings
18 5.5.3.1 Correlations of Personal and Indoor Concentrations with Ambient Concentrations
19 of Particulate Matter Constituents
20 The elemental composition of PM in personal samples was measured in the PTEAM study,
21 the first probability-based study of personal exposure to particles. A number of important
22 observations, made from the PTEAM data collected in Riverside, CA, are summarized by
23 Ozkaynak et al. (1996b). Population-weighted daytime personal exposures averaged
24 150 ± 9 //g/m3, compared to concurrent indoor and outdoor concentrations of 95 ± 6 //g/m3. The
25 personal exposure measurements suggested that there was a "personal cloud" of particles
26 associated with personal activities. Daytime personal exposures to 14 of the 15 elements
27 measured in the samples were considerably greater than concurrent indoor or outdoor
28 concentrations, with sulfur being the only exception.
29 The PTEAM data also showed good agreement between the concentrations of the elements
30 measured outdoors at the backyard of the residences with the concentrations measured at the
31 central site in the community. The agreement was excellent for sulfur. Although the particle and
April 2002 5-77 DRAFT-DO NOT QUOTE OR CITE
-------
1 element mass concentrations were higher in personal samples than for indoor or outdoor samples,
2 a nonlinear mass-balance method showed that the penetration factor was nearly 1 for all particles
3 and elements.
4 Similarly to the PTEAM results, recent measurements of element concentrations in
5 NHEXAS showed elevated concentrations of As and Pb in personal samples relative to indoor
6 and outdoor samples (Clayton et al., 1999b). The elevated concentrations of As and Pb were
7 consistent with elevated levels of PM in personal samples (median particle exposure of
8 101 //g/m3), compared to indoor concentrations (34.4 //g/m3). There was a strong association
9 between personal and indoor concentrations and indoor and outdoor concentrations for both As
10 and Pb. However, there were no central site ambient measurements for comparison to the
11 outdoor or indoor measurements at the residences.
12 Manganese (Mn) concentrations were measured in PM2 5 samples collected in Toronto
13 (Crump, 2000). The mean PM25 Mn concentrations were higher outdoors than indoors. But the
14 outdoor concentrations measured at the participant's homes were lower than those measured at
15 two fixed locations. Crump (2000) suggested that the difference in the concentrations may have
16 been because the fixed locations were likely closer to high-traffic areas than were the
17 participant's homes.
18 Studies of acidic aerosols and gases typically measure strong acidity (H+), SO42", NH4+, and
19 NO3". The relationship between the concentrations of these ions and the relationship between
20 indoor and outdoor concentrations have been addressed in a number of studies during which
21 personal samples, microenvironmental, and outdoor samples have been collected, as shown in
22 Tables 5-14 and 5-15. Key findings from these studies include those shown below.
23 • Acid aerosol concentrations measured at the residences in the Uniontown, PA, study were
24 significantly different from those measured at a fixed ambient site located 16 km from the
25 community. But, Leaderer et al. (1999) reported that the regional ambient air monitoring
26 site in Vinton, VA, provided a reasonable estimate of indoor and outdoor sulfate
27 measurements during the summer at homes without tobacco combustion.
28 • Approximately 75% of the fine aerosol indoors during the summer was associated with
29 outdoor sources based on I/O sulfate ratios measured in the Leaderer et al. (1999) study.
30 • Personal exposures to strong acidity (H+) were lower than corresponding outdoor levels
31 measured in studies by Brauer et al. (1989, 1990) and Suh et al. (1992). But the personal
April 2002 5-78 DRAFT-DO NOT QUOTE OR CITE
-------
>
to
o
o
to
TABLE 5-14. STUDIES THAT HAVE MEASURED PARTICULATE MATTER CONSTITUENTS IN PERSONAL
EXPOSURE SAMPLES
PM Constituent Study Name/Reference
Study Location
Population Size/No, of Samples
Summary of Results
vo
Elements
PTEAM/Ozkaynak et al. (1996b) Riverside, CA
178 adults
As and Pb
Mn
Acid Aerosol
Constituents
NHEXAS/Clayton et al. (1999b) EPA Region 5
Pellizzarietal. (1998, 1999)
Clayton etal. (1999a),
Crump (2000)
Samat et al. (2000)
Braueretal. (1989)
Suhetal. (1992)
Toronto
Baltimore, MD
Boston, MA
Uniontown, PA
167 samples
925 personal samples
20 adults
24 children for 2 days
Outdoor air was the major source for most elements
indoors, providing 70 to 100% of the observed indoor
concentrations for 12 of the 15 elements. Correlation
coefficients for central monitoring site versus outdoor at
the residences were 0.98 for sulfur and 0.5 to 0.9 for other
elements (except copper).
Personal As and Pb levels higher than indoor or outdoor
levels. No community ambient site for comparison.
Mean PM2 5 Mn higher outdoors than indoors. But PM2 5
Mn concentrations higher at two fixed locations than at
participants' homes.
High correlations between personal and ambient sulfate
measurements in summer and winter.
Personal exposures to aerosol strong acidity slightly lower
than concentrations measured at stationary site.
Personal exposures to H+ and SO4"2 lower than outdoor
levels, but higher than indoor microenvironmental levels;
personal NH4+ and NO3" higher than indoor or outdoor
levels.
H
6
o
0
H
O
O PAHs
s
O
H
W
Suhetal. (1993a,b)
Suhetal. (1994)
Waldman and Liang (1993),
Waldman etal. (1990)
Zmirou et al. (2000)
State College, PA 47 children
Georgia and New Hospital, daycares
Jersey
Grenoble, France 38 adults
Results similar to Uniontown, PA, study.
Results indicate strong neutralization of acidity indoors.
Indoor sulfate levels were 70 to 100% of outdoor levels.
Indoor ammonia levels 5- to 50-times higher than
outdoors. Indoors, acid aerosols were largely neutralized.
Ambient air concentrations close to traffic emissions were
1 . 1 - to 3 . 5 - times higher than personal exposure
concentrations.
-------
>
to
o
o
to
TABLE 5-15. STUDIES THAT HAVE MEASURED PARTICULATE MATTER CONSTITUENTS IN
MICROENVIRONMENTAL SAMPLES
PM Constituent Study Name/Reference
Study Location
Population Size/No, of Samples
Summary of Results
oo
O
H
6
o
o
H
O
^w'
s
Acid Aerosol Jones et al. (2000)
Constituents
PAHs
PAHs and
phthalates
Birmingham, England 12 residences
Patterson and Eatough (2000) Lindon, UT One school
Leaderer et al. (1999)
Braueretal. (1990)
Chuang et al. (1999)
Dubowsky et al. (1999)
Sheldon et al. ( 1 993a,b)
PTEAM/Ozkaynak et al.
(1996b),
Sheldon etal. (1993c)
Virginia and
Connecticut
Boston, MA
Durham, NC
Boston, MA
Placerville and
Ro Seville, CA
Riverside, CA
232 homes
11 homes
24 homes
3 buildings
280 homes
120 homes
Sulfate I/O ratios ranged from 0.7 to 0.9 for three PM size
fractions.
Ambient sulfate, SO2, nitrate, soot, and total particle
number showed strong correlations with indoor exposure,
although ambient PM2 5 mass was not a good indicator of
total PM25 exposure.
The regional ambient air monitoring site provided a
reasonable estimate of indoor and outdoor sulfate at
nonsmokers homes. I/O sulfate ratio of 0. 74 during
summer. Ammonia concentrations were an order of
magnitude higher indoors than outdoors. Nitrous acid
levels higher indoors than outdoors.
Outdoor levels of H+, SO2, HNO3, and SO4'2 exceeded
indoor levels in winter and summer. I/O ratios of H+
lower than I/O ratios of SO4"2 indicated neutralization of
the acidity by ammonia.
Measurements with continuous monitor; PAH levels
generally higher indoors than outdoors.
PAHs indoors attributable to traffic, cooking, and candle-
burning.
Mass balance model used to estimate source strengths for
PAH sources such as smoking, wood-burning and
cooking.
12-h I/O ratios for particulate-phase PAHs ranged from
1.1 to 1.4 during the day and 0.64 to 0.85 during night.
The concentrations of phthalates and the number of
samples with detectable phthalates were higher indoors
than outdoors.
O
HH
H
W
-------
1 exposure levels measured by Suh et al. (1992) were higher than the indoor
2 microenvironmental levels.
3 • Personal exposures to NH4+, and NO3" were reported by Suh et al. (1992) to be lower than
4 either indoor or outdoor levels.
5 • Personal exposures to SO42" were also lower than corresponding outdoor levels, but
6 higher than the indoor microenvironmental levels (Suh et al., 1992; 1993a,b), as shown in
7 Table 5-16.
8
9 The fact that the personal and indoor H+ concentrations were substantially lower than
10 outdoor concentrations suggests that a large fraction of aerosol strong acidity is neutralized by
11 ammonia. Ammonia is emitted in relatively high concentrations in exhaled breath and sweat.
12 The difference between indoor and outdoor H+ concentrations in the Suh et al. (1992, 1993a,b)
13 studies was also much higher than the difference for indoor and outdoor SO42", indicative of
14 neutralization of the H+. Results of the Suh et al. (1992, 1993a,b) studies also showed substantial
15 interpersonal variability of H+ concentrations that could not be explained by variation in outdoor
16 concentrations.
17 Similar results for ammonia were reported by Waldman and Liang (1993). They reported
18 that levels of ammonia in institutional settings that they monitored were 10- to 50- times higher
19 than outdoors, and that acid aerosols were largely neutralized. Leaderer et al. (1999) reported
20 that ammonia concentrations during both winter and summer in residences were an order of
21 magnitude higher indoors than outdoors, consistent with results of other studies and the presence
22 of sources of ammonia indoors.
23 Sulfate aerosols appear to penetrate indoors effectively. Waldman et al. (1990) reported
24 I/O ratios of 0.7 to 0.9 in two nursing care facilities and a day-care center. Sulfate I/O ratios were
25 measured for three particle size fractions in 12 residences in Birmingham, England, by Jones
26 et al. (2000). The sulfate I/O ratios were 0.7 to 0.9 for PM< 1.1 //m, 0.6 to 0.8 for PM 1.1 to
27 2.1 //m, and 0.7 to 0.8 for PM 2.1 to 10 //m. Suh et al. (1993b) reported that personal and
28 outdoor sulfate concentrations were highly correlated, as depicted in Figure 5-10.
29 Indoor/outdoor relationships were measured for a number of PM25 components and related
30 species in Lindon, UT, during January and February of 1997 by Patterson and Eatough (2000).
31 Outdoor samples were collected at the Utah State Air Quality monitoring site. Indoor samples
April 2002 5-81 DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-16. SUMMARY STATISTICS FOR PERSONAL, INDOOR, AND
OUTDOOR CONCENTRATIONS OF SELECTED AEROSOL COMPONENTS IN
TWO PENNSYLVANIA COMMUNITIES
Concentration (nmol m"3)
Aerosol
State College
NO3
so42-
NH4+
H+
Uniontown
S042'
NH4+
H+
Home Type
A/C Homes0
Non-A/C
A/C Homes
Non-A/C
All Homes'1
All Homes
A/C Homes
Non-A/C
All Homes6
All Homes6
All Homes6
All Homes6
Sample Site
(In/Out)a
53/71
254/71
56/75
259/75
214/76
314/155
28/74
230/74
163/75
91/46
91/44
91/46
Indoor (12 h)
GM ± GSDb
2.1 ±2.7
3.2 ±2.3
61.8 ±2.5
96.7 ±2.5
69.1 ±2.6
154.7 ±2.8
4.2 ±4.3
11.2±3.1
9.1 ±3. 5
87.8 ±2.1
157.2 ±2.8
13.7 ±2.5
Outdoor (24 h)
GM ± GSDb
1.4 ±2.1
1.4 ±2.1
109.4 ±2.4
109.4 ±2.4
91.0 ±2.5
104.4 ±2.3
82.5 ±2.6
82.5 ±2.6
72.4 ±2.9
124.9 ±1.9
139.4 ±2.1
76.6 ±2.7
Personal (12 h)
GM ± GSDb
—
—
71.5 ±2.4
—
18.4 ±3.0
110.3 ± 1.8
167.0 ±2.0
42.8 ±2.2
aIn/Out = Indoor sample site/outdoor sample site.
bGM ± GSD = Geometric mean ± geometric standard deviation.
°A/C Homes = Homes that had air-conditioning (A/C); this does not imply that it was on during the entire
sampling period.
Non-A/C = Homes without air conditioning.
dThe sample size (n) for the personal monitoring = 209.
6n = 174 for personal monitoring.
Source: Suhetal. (1992, 1993a,b).
1 were collected in the adjacent Lindon Elementary School. The infiltration factors, Cai/Cao, given
2 by the slope of the regression lines (Table 5-17), were low (0.27 for sulfate and 0.12 for PM2 5),
3 possibly because of removal of particles in the air heating and ventilation system. The authors
4 concluded that the data indicate that indoor PM2 5 mass may not always be a good indicator of
5 exposure to ambient combustion material caused by the influence of indoor sources of particles.
6 However, ambient sulfate, SO2, nitrate, soot, and total paniculate number displayed strong
7 correlations with indoor exposure. Ambient PM2 5 mass was not a good indicator of indoor PM2 5
8 mass exposure.
April 2002
5-82
DRAFT-DO NOT QUOTE OR CITE
-------
600
600
Outdoor Sulfate (nmoles/rrr)
Figure 5-10. Personal versus outdoor SO4 in State College, PA. Open circles represent
children living in air conditioned homes; the solid line is the 1:1 line.
Source: Suhetal. (1993b).
1 Oglesby et al. (2000) conducted a study to evaluate the validity of fixed-site fine particle
2 concentration measurements as exposure surrogates for air pollution epidemiology. Using 48-h
3 EXPOLIS data from Basel, Switzerland, they investigated the personal exposure/outdoor
4 concentration relationships for four indicator groups: (1) PM25 mass, (2) sulfur and potassium
5 for regional air pollution, (3) lead and bromine for traffic-related particles, and (4) calcium for
6 crustal particles. The authors reported that personal exposures to PM2 5 mass were not correlated
April 2002
5-83
DRAFT-DO NOT QUOTE OR CITE
-------
TABLE 5-17. STATISTICAL CORRELATION OF OUTDOOR (x) VERSUS INDOOR
(y) CONCENTRATION FOR MEASURED SPECIES
(Units are nmol m3, except for soot and metals, which are
and absorption units m"3, respectively.)"
Species
SO2 All Samples
SO2 Day Samples
SO2 Night Samples
Sulfate All Samples
Sulfate Day Samples
Sulfate Night Samples
Nitrate All Samples
Nitrate Day Samples
Nitrate Night Samples
Soot Day Samples
Soot Night Samples
Total Acidity All Samples
Metals All Samples
Slope
0.0272 ± 0.0023
0.0233 ± 0.0037
0.0297 ± 0.0029
0.267 ± 0.024
0.261 ±0.034
0.282 ±0.035
0.0639 ±0.0096
0.097 ±0.0096
0.047 ±0.011
0.43 ± 0.25
0.33 ±0.13
0.04 ±0.73
0.10±0.30
Intercept
0.34±0.13
0.75 ±0.26
0.099 ±0.075
-0.14 ±0.48
0.40 ±0.66
-0.84 ±0.68
0.9± 1.5
-0.4 ± 1.4
1.5 ± 1.8
3.5 ± 1.7
0.00 ±0.55
0.42 ±0.23
0.0014 ±0.0042
r2
0.73
0.62
0.82
0.70
0.71
0.70
0.54
0.88
0.44
0.43
0.69
0.00
0.01
Average
Outdoors
38
56
20
16
16
16
134
126
139
6
4
0.2
0.0042
"Lindon Elementary School, Lindon, UT, January and February 1997.
Source: Patterson and Eatough (2000).
1 to corresponding home outdoor levels (n = 44, r = 0.07). In the study group reporting neither
2 relevant indoor sources nor relevant activities, personal exposures and home outdoor levels of
3 sulfur were highly correlated (n = 40, r = 0.85). These results are consistent with spatially
4 homogeneous regional pollution and higher spatial variability of traffic and crustal materials.
5 PAHs have been measured in studies by EPA and the California Air Resources Board.
6 PAH results from a probability sample of 125 homes in Riverside are discussed in reports by
7 Sheldon et al. (1992a,b) and Ozkaynak et al. (1996b). Data for two sequential 12-h samples were
8 reported for PAHs by ring size (3 to 7) and for individual phthalates. The results are summarized
9 below.
10 • The particulate-phase 5- to 7-ring species had lower relative concentrations than the more
11 volatile 3- to 4-ring species.
April 2002 5-84 DRAFT-DO NOT QUOTE OR CITE
-------
1 • The 12-h indoor/outdoor ratios for the 5- to 7-ring species ranged from 1.1 to 1.4 during
2 the day and from 0.64 to 0.85 during the night (Sheldon et al., 1993a).
3 • An indoor air model used to calculate indoor "source strengths" for the PAHs showed
4 that smoking had the strongest effect on indoor concentrations.
5 Results from a larger PAH probability study in 280 homes in Placerville and Roseville
6 (Sheldon et al., 1993a,b) were similar to the 125-home study. The higher-ring, particle-bound
7 PAH's had lower indoor and outdoor concentrations than the lower-ring species. For most
8 PAHs, the I/O ratio was greater than 1 for smoking and smoking/fireplace homes and less than
9 1 for fireplace-only, wood stove, wood stove/gas heat, gas heat, and "no source" homes.
10 A study of PAHs in indoor and outdoor air was conducted in 14 inner-city and 10 rural
11 low-income homes near Durham, NC, in two seasons (winter and summer) in 1995 (Chuang
12 etal., 1999). Fine-particle-bound PAH concentrations measured with a real-time monitor were
13 usually higher indoors than outdoors (2.47 ± 1.90 versus 0.53 ± 0.58 //g/m3). Higher indoor
14 levels were seen in smoker's homes compared with nonsmoker's homes, and higher outdoor and
15 indoor PAH levels were seen in urban areas compared with rural areas.
16 In a study reported by Dubowsky et al. (1999), the weekday indoor PAH concentrations
17 attributable to traffic (indoor source contributions were removed) were 39 ± 25 ng/m3 in a
18 dormitory that had a high air exchange rate because of open windows and doors, 26 ± 25 ng/m3
19 in an apartment, and 9 ± 6 ng/m3 in a suburban home. The study showed that both
20 outdoor—especially motor vehicular traffic—and indoor sources contributed to indoor PAH
21 concentrations. BaP concentrations were measured in the THEES study (Waldman et al., 1991).
22 A comprehensive analysis of the data showed considerable seasonal variability of indoor and
23 outdoor sources and resultant changes in personal exposures to BaP.
24
25 5.5.4 Factors Affecting Correlations Between Ambient Measurements and
26 Personal or Microenvironmental Measurements of Particulate Matter
27 Constituents
28 The primary factors affecting correlations between personal exposure and ambient air PM
29 measurements have been discussed in Section 4.3.2. These include air-exchange rates, particle
30 penetration factors, decay rates and removal mechanisms, indoor air chemistry, indoor sources,
31 and freshly-generated particles indoors. The importance of these factors varies for different PM
April 2002 5-85 DRAFT-DO NOT QUOTE OR CITE
-------
1 constituents. For acid aerosols, indoor air chemistry is particularly important as indicated by the
2 discussion of the neutralization of the acidity by ammonia, which is present at higher
3 concentrations indoors because of the presence of indoor sources. For SVOCs, including PAHs
4 and phthalates, the presence of indoor sources will impact substantially the correlation between
5 indoor and ambient concentrations (Ozkaynak et al., 1996b; Sheldon et al., 1993b). Penetration
6 factors for PM will impact correlations between indoors and outdoors for most elements, except
7 Pb, which may have significant indoor sources in older homes. Indoor air chemistry, decay rates,
8 and removal mechanisms may affect soot and organic carbon. Furthermore, reactions between
9 indoor and outdoor gases and particles may also produce freshly generated aerosols indoors.
10 These factors must be fully evaluated when attempting to correlate ambient, personal, and indoor
11 PM concentrations.
12
13 5.5.5 Limitations of Available Data
14 The previous discussion demonstrates that there is very limited data available that can be
15 used to compare personal, microenvironmental, and ambient air concentrations of PM
16 constituents. Because of resource limitations, PM constituents have not been measured in many
17 studies of PM exposure. There are little data on freshly generated aerosols indoors. Although
18 there is some data on acid aerosols, the comparisons between the personal and indoor data
19 generally have been with outdoor measurements at the participant's residences, not with
20 community ambient air measurement sites. The relationship between personal exposure and
21 indoor levels of acid aerosols is not clear because of the limited database. The exception is
22 sulfate, for which there appears to be a strong correlation between indoor and ambient
23 concentrations.
24 With the exception of PAHs, there are practically no data available to relate personal or
25 indoor concentrations with outdoor or ambient site concentrations of SVOCs, which may be
26 generated from a variety of combustion and industrial sources. The relationship between
27 exposure and ambient concentrations of particles from specific sources, such as diesel engines,
28 has not been determined.
29 Although there is an increasing amount of research being performed to measure PM
30 constituents in different PM size fractions, the current data are inadequate to adequately assess
31 the relationship between indoor and ambient concentrations of most PM constituents. Another
April 2002 5-86 DRAFT-DO NOT QUOTE OR CITE
-------
1 area where additional information has to be developed is the PM exposures that are derived from
2 outdoor vapors (ov) reacting (rxn) with indoor vapors (iv). This is a source that could also vary
3 with outdoor PM, for example, when the (ov) is ozone.
4
5
6 5.6 IMPLICATIONS OF USING AMBIENT PARTICULATE MATTER
7 CONCENTRATIONS IN EPIDEMIOLOGIC STUDIES OF
8 PARTICULATE MATTER HEALTH EFFECTS
9 In this section, the exposure issues that relate to the interpretation of the findings from
10 epidemiologic studies of PM health effects are examined. This section examines the errors that
11 may be associated with using ambient PM concentrations in epidemiologic analyses of PM health
12 effects. First, implications of associations found between personal exposure and ambient PM
13 concentrations are reviewed. This is discussed separately in the context of either community
14 time-series studies or long-term, cross-sectional studies of chronic effects. Next, the role of
15 compositional and spatial differences in PM concentrations are discussed and how these may
16 influence the interpretation of findings from PM epidemiology. Finally, using statistical
17 methods, an evaluation of the influence of exposure measurement errors on PM epidemiology
18 studies is presented.
19
20 5.6.1 Potential Sources of Error Resulting from Using Ambient Particulate
21 Matter Concentrations in Epidemiologic Analyses
22 Measurement studies of personal exposures to PM are still few and limited in spatial,
23 temporal, and demographic coverage. Consequently, with the exception of a few longitudinal
24 panel studies, most epidemiologic studies of PM health effects rely on ambient community
25 monitoring data giving 24-h average PM concentration measurements. Moreover, because of
26 limited sampling for PM2 5, many of these epidemiologic studies had to use available PM10 or in
27 some instances had to rely on historic data on other PM measures or indicators, such as TSP,
28 SO4=, TP15, RSP, COH, KM, etc. A critical question often raised in the interpretation of results
29 from acute or chronic epidemiologic community-based studies of PM is whether the use of
30 ambient stationary site PM concentration data influences or biases the findings from these
31 studies. Because the health outcomes are measured on individuals, the epidemiologists might
April 2002 5-87 DRAFT-DO NOT QUOTE OR CITE
-------
1 prefer to use personal exposure measurements (total, ambient, or nonambient) instead of
2 surrogates, such as ambient PM concentration measurements collected at one or more ambient
3 monitoring sites in the community. Use of ambient concentrations could lead to
4 misclassification of individual exposures and to errors in the epidemiologic analysis of pollution
5 and health data depending on the pollutant and on the mobility and lifestyles of the population
6 studied. Ambient monitoring stations can be some distance away from the individuals and can
7 represent only a fraction of all likely outdoor microenvironments that individuals come in contact
8 with during the course of their daily lives. Furthermore, most individuals are quite mobile and
9 move through multiple microenvironments (e.g., home, school, office, commuting, shopping,
10 etc.) and engage in diverse personal activities at home (e.g, cooking, gardening, cleaning,
11 smoking). Some of these microenvironments and activities may have different sources of PM
12 and result in distinctly different concentrations of PM than that monitored by the fixed-site
13 ambient monitors. Consequently, exposures of some individuals will be classified incorrectly if
14 only ambient monitoring data are used to estimate individual level exposures to PM. Thus, bias
15 or loss of precision in the epidemiologic analysis may result from improper assessment of
16 exposures using data routinely collected by the neighborhood monitoring stations.
17 Because individuals are exposed to particles in a multitude of indoor and outdoor
18 microenvironments during the course of a day, concern over error introduced in the estimation of
19 PM risk coefficients using ambient, as opposed to personal, PM measurements has received
20 considerable attention recently from exposure analysts, epidemiologists, and biostatisticians.
21 Some exposure analysts contend that, for community time-series epidemiology to yield
22 information on the statistical association of a pollutant with a health response, there must be an
23 association between personal exposure to a pollutant and the ambient concentration of that
24 pollutant because people tend to spend around 90% time indoors and are exposed to both indoor
25 and outdoor-generated PM (cf. Wallace, 2000b; Brown and Paxton, 1998; Ebelt et al., 2000).
26 Consequently, numerous findings reported in the epidemiologic literature on significant
27 associations between ambient PM concentrations and various morbidity and mortality health
28 indices, in spite of the low correlations between ambient PM and concentrations and measures of
29 personal exposure, has been described by some exposure analysts as an exposure paradox
30 (Lachenmyer and Hidy, 2000, Wilson et al., 2000).
April 2002 5-88 DRAFT-DO NOT QUOTE OR CITE
-------
1 To resolve the so-called exposure paradox, several types of analyses need to be considered.
2 The first type of analysis has to examine the correlations between ambient PM concentrations
3 and personal exposures that are relevant to most of the existing PM epidemiology studies using
4 either pooled, daily-average, or longitudinal exposure data. The second approach has to study the
5 degree of correlations between the two key components of personal PM exposures (i.e.,
6 exposures caused by ambient-generated PM and exposures caused by nonambient PM) with
7 ambient or outdoor PM concentrations, for each of the three types of exposure study design. Yet,
8 even with these two approaches, it may still be difficult to examine complex synergisms which,
9 in some situations, may preclude simple decoupling of in indoor and outdoor particles either in
10 terms of exposure or total dose delivered to the lung. In addition, several factors influencing
11 either the exposure or health response characterization of the subjects have to be addressed.
12 These include such factors as:
13 • spatial variability of PM components,
14 • health or sensitivity status of subjects,
15 • variations of PM with other co-pollutants,
16 • co-generation of fine and ultrafine particles from outdoor air and indoor gaseous
17 pollutants,
18 • formal evaluation of exposure errors in the analysis of health data, and
19 • how the results may depend on the variations in the design of the epidemiologic study.
20 To facilitate the discussion of these topics, a brief review of concepts pertinent to exposure
21 analysis issues in epidemiology is presented.
22
23 5.6.2 Associations Between Personal Exposures and Ambient Particulate
24 Matter Concentrations
25 As defined earlier in Sections 5.3 and 5.4, personal exposures to PM result from an
26 individual's exposures to PM in many different types of microenvironments (e.g., outdoors near
27 home, outdoors away from home, indoors at home, indoors at office or school, commuting,
28 restaurants, malls, other public places, etc.). Total personal exposures (Et) that occur in these
29 indoor and outdoor microenvironments can be classified as those resulting from PM of outdoor
30 origin (Eag ) and those primarily generated by indoor sources and personal activities (Enonag =
31 Epig+Epact). The associations between personal exposures and ambient PM concentrations that
April 2002 5-89 DRAFT-DO NOT QUOTE OR CITE
-------
1 have been reported from various personal exposure monitoring studies under three broad
2 categories of study design: (1) longitudinal, (2) daily-average, or (3) pooled exposure studies are
3 summarized below.
4 In the previous Sections 5.4.3.1.2 and 5.4.3.1.3, some recent studies mainly conducted in
5 the United States, and involving children, the elderly, and subjects with COPD were reviewed,
6 and they indicated that both intra- and interindividual variability in the relationships between
7 personal exposures and ambient PM concentrations were observed. A variety of different
8 physical, chemical, and personal or behavioral factors were identified by the original
9 investigators that seem to influence the magnitude and the strength of the associations reported.
10 Clearly, for cohort studies in which individual daily health response are obtained,
11 individual longitudinal PM personal exposure data (including ambient-generated and nonambient
12 components) provide the appropriate indicators. In this case, health responses of each individual
13 can be associated with the total personal exposure, the ambient-generated exposure, or the
14 nonambient exposure of each individual. Also, the relationships of personal exposure indicators
15 with ambient concentration can be investigated. In the case of community time-series
16 epidemiology, however, it is not feasible to obtain experimental measurements of personal
17 exposure for the millions of people over time periods of years that are needed to investigate the
18 relationship between air pollution and infrequent health responses such as deaths or even hospital
19 admissions. The epidemiologist must work with the aggregate number of health responses
20 occurring each day and a measure of the ambient concentration that is presumed to be
21 representative of the entire community. The relationship of PM exposures of the potentially
22 susceptible groups to monitored ambient PM concentrations depends on their activity pattern and
23 level, residential building and HVAC factors (which influence the infiltration factor), status of
24 exposure to ETS, amount of cooking or cleaning indoors, and seasonal factors, among others.
25 Average personal exposures of these special subgroups to ambient-generated PM are correlated
26 well with ambient PM concentrations regardless of individual variation in the absence of major
27 microenvironmental sources.
28 Even though both Eag and Enonag contributes to daily baseline PM dose received by the lung,
29 there seem to be clear differences in the relationships of ambient (Eag) and nonambient (Enonag)
30 exposure with ambient concentration (Ca). Various researchers have shown that Enonag is
31 independent of Ca, but that Eag is a function of Ca. Wilson et al. (2000) explains the difference
April 2002 5-90 DRAFT-DO NOT QUOTE OR CITE
-------
1 based on different temporal patterns that affect PM concentrations. "Concentrations of ambient
2 PM are driven by meteorology and by changes in the emission rates and locations of emission
3 sources, while concentrations of nonambient PM are driven by the daily activities of people."
4 Still, although Enonag may not correlate with ambient Ca or Eag, it will nevertheless add to the daily
5 baseline dose received by the lung.
6 Ott et al. (2000) also discuss the reasons for assuming that Enonag is independent of Eag and
7 Ca. They show that the nonambient component of total personal exposure is uncorrelated with
8 the outdoor concentration data. Ott et al. (2000) show the Eno is similar for three population-
9 based exposure studies, including two large probability-based studies, the PTEAM study
10 conducted in Riverside (Clayton et al., 1993; Thomas et al., 1993; Ozkaynak et al., 1996a,b) and
11 a study in Toronto (Pelizzarri et al., 1999; Clayton et al., 1999a), as well as a nonprobability-
12 based study, conducted in Phillipsburg (Lioy et al., 1990). Based on these three studies, they
13 conclude that Enonag and the distribution of(Enonag)it can be treated as constant from city to city,
14 where i refers to a specific individual and t to a specific day..
15 Dominici et al. (2000) examined a larger database consisting of five different PM exposure
16 studies and concluded that Eno can be treated as relatively constant from city to city.
17 If (Enonag)t were constant, this would imply that it would have a zero correlation with (Ca)t.
18 However, this hypothesis of constant (Enonag)it has not been established fully because only a few
19 studies have obtained the data needed to estimate (Enonag)it. Although Eno is independent of
20 Ca, it may not be independent of a. Sarnat et al. (2000) show that Eno goes up as the
21 ventilation rate (and a) goes down. Lachenmeyer and Hidy (2000) also show, by comparing
22 winter and summer regression equations, that as the slope (a) goes down, the intercept (Enonag)
23 goes up.
24 Mage et al. (1999) assume that the PM10 concentration component from indoor sources
25 (such as smoking, cooking, cleaning, burning candles, and so on) is not correlated with the
26 outdoor concentration. They indicate that this lack of correlation is expected, because people are
27 unaware of ambient concentrations and do not necessarily change their smoking or cooking
28 activities as outdoor PM10 concentrations vary, an assumption supported by other empirical
29 analyses of personal exposure data. For the PTEAM data set, Mage et al. (1999) have shown that
30 Epig and Ca have r near zero (R2 = 0.005). Wilson et al. (2000) have shown the Cai and Cpig also
April 2002 5-91 DRAFT-DO NOT QUOTE OR CITE
-------
1
2
3
4
have r near zero (R2 = 0.03). Figure 5-11 shows the relationship of estimated (Enonag)it and En,
with Ca (calculated by EPA from PTEAM and THEES data).
ionag
200
Cl)
E 150-
C/5 O)
O 13-
Q. _
IN "= 10°
_ .9^
CO ^
m /—
Q- o
-50
(a)
* *
A *
**»
* * *
* +«.**+ *
50
100
150
200
250
60-
o _-
£ c
0) O>
Q- J2
co
50-
40-
30-
O
Q.
X
LLJ 20-
(b)
* *
50 100 150
Ambient Concentration,
200
Figure 5-11. Plots of nonambient exposure to PM10, (a) daytime individual values from
PTEAM data and (b) daily-average values from THEES data.
Source: Data taken from (a) Clayton et al. (1993) and (b) Lioy et al. (1990).
April 2002
5-92
DRAFT-DO NOT QUOTE OR CITE
-------
1 Based on these results it is reasonable to assume that ordinarily Enonag has no relationship
2 with Ca in the absence of sufficient study-specific data on the complex interactions between
3 indoor and outdoor gases and particles producing fresh particles indoors. Therefore, in linear
4 nonthreshold models of PM health effects, Enonag is not expected to contribute to the relative risk
5 determined in a regression of health responses on Ca. Furthermore, in time-series analysis of
6 pooled or daily heath data, it is expected that Eag rather than Et will have the stronger association
7 with Ca.
8
9 5.6.3 Role of Compositional Differences in Exposure Characterization
10 for Epidemiology
11 The majority of the available data on PM exposures and relationships with ambient PM
12 have come from a few large-scale studies, such as PTEAM, or longitudinal studies on selected
13 populations, mostly the elderly. Consequently, for most analyses, exposure scientists and
14 statisticians had to rely on PM10 or PM2 5 mass data, instead of elemental or chemical
15 compositional information on individual or microenvironmental samples. In a few cases,
16 researchers have examined the factors influencing indoor outdoor ratios or penetration and
17 deposition coefficients using elemental mass data on personal, indoor, and outdoor PM data (e.g.,
18 Ozkaynak et al. 1996a,b; Yakovleva et al. 1999). These results have been informative in terms
19 of understanding relative infiltration of different classes of particle sizes and sources into
20 residences (e.g., fossil fuel combustion, mobile source emissions, soil-derived, etc.). Clearly, in
21 the accumulation-mode, particles associated with stationary or mobile combustion sources have
22 greater potential for penetration into homes and other microenvironments than do crustal
23 material. The chemical composition of even these broad categories of source classes may have
24 distinct composition and relative toxicity. Moreover, when particles and reactive gases are
25 present indoors in the presence of other pollutants or household chemicals, they may react to
26 form additional or different compounds and particles with yet unknown physical, chemical, and
27 toxic composition (Wainman et al. 2000). Thus, if indoor-generated and outdoor-generated PM
28 were responsible for different types of health effects, or had significantly different toxicities on a
29 per unit mass basis, it would be then be important that Eag and Enonag should be separated and
30 treated as different species, much like the current separation of PM10 into PM25 and PM10_25.
31 These complexities in personal exposure profiles may introduce nonlinearities and other
April 2002 5-93 DRAFT-DO NOT QUOTE OR CITE
-------
1 statistical challenges in the selection and fitting of concentration-response models.
2 Unfortunately, PM health effects models have not yet been able to meaningfully consider such
3 complexities. The relationships of toxicity to the chemical and physical properties of PM are
4 discussed in Chapter 7.
5 It is important also to note that individuals spend time in places other than their homes and
6 outdoors. Many of the interpretations reported in the published literature on factors influencing
7 personal PM10 exposures, as well as in this chapter, come from the PTEAM study. The PTEAM
8 study was conducted 10 years ago in one geographic location in California, during one season,
9 and most residences had very high and relatively uniform air-exchange rates. Nonhome indoor
10 microenvironments were not monitored directly during the PTEAM study. Commuting
11 exposures from traffic or exposures in a variety of different public places or office buildings
12 could not be assessed directly. Nonresidential buildings may have lower or higher ambient
13 infiltration rates depending on the use and type of the mechanical ventilation systems employed.
14 Because the source and chemical composition of particulate matter effecting personal exposures
15 in different microenvironments vary by season, day-of-the-week, and time of day, it is likely that
16 some degree of misclassification of exposures to PM toxic agents of concern will be introduced
17 when health effects models use only daily-average mass measures such as PM10 or PM2 5.
18 Because of the paucity of currently available data on many of these factors, it is impossible to
19 ascertain at this point the magnitude and severity of these more complex exposure
20 missclassification problems in the interpretation of results from PM epidemiology.
21
22 5.6.4 Role of Spatial Variability in Exposure Characterization for
23 Epidemiology
24 Chapter 3 (Section 3.2.3) and Chapter 5 (Section 5.3) present information on the spatial
25 variability of PM mass and chemical components at fixed-site ambient monitors; for purposes of
26 this chapter, this spatial variability is called an "ambient gradient." Any gradient that may exist
27 between a fixed-site monitor and the outdoor microenvironments near where people live, work,
28 and play, obviously affects the concentration profile actually experienced by people as they go
29 about their daily lives.
30 However, the evidence so far indicates that PM concentrations, especially fine PM (mass
31 and sulfate), generally are distributed uniformly in most metropolitan areas. This reduces the
April 2002 5-94 DRAFT-DO NOT QUOTE OR CITE
-------
1 potential for exposure misclassification because of outdoor spatial gradients when a limited
2 number of ambient PM monitors are used to represent population average ambient exposures in
3 time-series or cross-sectional epidemiologic studies of PM. This topic is further discussed below
4 in Section 5.6.5. However, as discussed earlier, the same assumption is not necessarily true for
5 different components of PM, because source-specific and other spatially nonuniform pollutant
6 emissions could alter the spatial profile of individual PM components in a community.
7 For example, particulate and gaseous pollutants emitted from motor vehicles tend to be higher
8 near roadways and inside cars. Likewise, acidic and organic PM species may be location- and
9 time-dependent. Furthermore, human activities are complex, and if outdoor PM constituent
10 concentration profiles are either spatially or temporally variable, it is likely that exposure
11 misclassification errors could be introduced in the analysis of PM air pollution and health data.
12
13 5.6.5 Analysis of Exposure Measurement Error Issues in Particulate Matter
14 Epidemiology
15 The effects of exposure misclassification on relative risk estimates of disease using
16 classical 2x2 contingency design (i.e., exposed/nonexposed versus diseased/nondiseased) have
17 been studied extensively in the epidemiologic literature. It has been shown that the magnitude of
18 the exposure-disease association (e.g., relative risk) because of either misclassification of
19 exposure or disease alone (i.e., nondifferential misclassification) biases the effect results toward
20 the null, and differential misclassification (i.e, different magnitudes of disease misclassification
21 in exposed and nonexposed populations) can bias the effect measure toward or away from the
22 null value relative to the true measure of association (Shy et al., 1978; Gladen and Rogan, 1979;
23 Copeland et al., 1977; Ozkaynak et al., 1986). However, the extension of these results from
24 contingency analysis design to multivariate (e.g., log-linear regression, Poissson, logit) models
25 typically used in recent PM epidemiology has been more complicated. Recently, researchers
26 have developed a framework for analyzing measurement errors typically encountered in the
27 analysis of time-series mortality and morbidity effects from exposures to ambient PM (cf. Zeger
28 et al., 2000; Dominici et al., 2000; Samet et al., 2000). Some analysis in the context of cross-
29 sectional epidemiology have also been conducted (e.g. Navidi et al., 1999).
30 The appropriateness of using ambient PM concentration as an exposure metric in the
31 context of epidemiologic analysis of health effects associated with exposure to PM recently has
April 2002 5-95 DRAFT-DO NOT QUOTE OR CITE
-------
1 been examined by a number of investigators (cf. Zeger et al., 2000; Dominici et al., 2000; Navidi
2 et al., 1999; Ozkaynak and Spengler, 1996). In the following section, the error analysis model
3 framework developed in Zeger et al. (2000) will be discussed in the context of time-series
4 epidemiology. After which, issues and implications of exposure errors to findings from long-
5 term/chronic or cross-sectional epidemiology will be discussed briefly.
6
7 5.6.5.1 Analysis of Exposure Measurement Errors in Time-Series Studies
8 The discussion presented in this section is further examined in Chapter 8 under the context
9 of implications of exposure errors to results and interpretation of findings from PM
10 epidemiology. The discussion presented in this section also focuses more on the potential for
1 1 exposure misclassification biases on the estimated regression slopes rather than on the more
12 subtle issues, such as those dealing with "effect modification" discussed further in Chapter 8.
13 Zeger et al. (2000) provide a useful framework for analyzing exposure error in community
14 time-series epidemiology. This framework, coupled with results from recent exposure studies,
15 makes it possible to clarify some important questions regarding relationships among the three
16 aspects of personal exposure (1) total personal, (2) personal caused by ambient PM, and
17 (3) personal resulting from nonambient PM and ambient concentration. Consider the regression
18 of a health response (i.e., mortality rate on day t, Yt, against the ambient concentration of PM on
19 day t, Ct). In analyzing pollution-level data on mortality and air pollution, log-linear regressions
20 of the form:
21
22
23 are fit, where Yt is the expected mortality rate; s(t) is an arbitrary but smooth function of time,
24 introduced to control for the confounding of longer trends and seasonality; Ct, is the average of
25 multiple monitor measurements of ambient pollution measurement for day t; and ut are other
26 possible confounders such as temperature and dew point on the same or previous day. Each
27 coefficient, P, in Equation 5-12 gives the expected change in the health response, Y, because of a
28 unit change in its corresponding variable.
April 2002 5-96 DRAFT-DO NOT QUOTE OR CITE
-------
1 However, instead of Equation 5-12, Zeger et al. (2000) suggest that the analyst would like
2 to know the corresponding relationship for personal exposure rather than ambient concentration,
3
Yt = exp[s(0 + EtpE + i*,AJ . (5-13)
4
5 Zeger et al. (2000) do not differentiate among the three aspects of personal or community
6 exposure. To understand the error in P caused by using ambient concentrations instead of
7 personal exposure in the regression analysis, it is necessary to examine the relationship between
8 Pc, based on a unit change in the ambient concentration, C, and PE, based on a unit change in one
9 of the three aspects of personal exposure, E. In considering the consequences for Pc, as an
10 estimate of PE, of having a measure of ambient pollution Ct, rather than actual personal exposure
1 1 Eit, it is convenient to express the desired pollution measurement, Eit, as Ct plus three error terms:
12
13 Ett = Ct + (Ett-Et)+(Et-Ct>(Ct*-Ct). (5-14)
14
15 Here Et represents the daily, community-average personal exposure. The first term,
16 (Eit - Et ), is the error resulting from having only aggregated or community-averaged exposure
17 rather than individual -level exposure data. The second term, (Et - Ct ), is the difference
18 between the average personal exposure and the true ambient pollutant level, and the third term,
19 (C*t -Ct), represents the difference between the true and the measured ambient concentration.
20 In the evaluation of these error terms, two types of measurement error often are considered
21 in the context of epidemiology. The classical error model assumes that measurement error,
22 (Ct-Et), depends on ambient measurements [simply referred to as Ct here instead of (Ca),]. The
23 Berkson error model assumes that the measurement error is dependent on the true value or the
24 personal exposure (Et). The regression coefficient (Pc), estimated from the health effects model
25 in the Berkson error case, gives an unbiased estimate of PE. In the classical error case, Pc is a
26 biased estimate of PE, and the degree of bias depends on the correlation between the
27 measurement error and Ct . The measurement error analysis of Zeger et al. (2000) includes three
28 components: (1) an individual's deviation from the risk-weighted average personal exposure;
April 2002 5-97 DRAFT-DO NOT QUOTE OR CITE
-------
1 (2) the difference between the average personal exposure and the true ambient level; and (3) the
2 difference between the measured and the true ambient levels, which include the spatial variation
3 of outdoor PM and instrument sampling error. Zeger et al. (2000) conclude that the first and
4 third components are of the Berkson type and, therefore, are likely to have smaller effects on the
5 relative risk estimates for PM. However, the second component can be a source of substantial
6 bias if, for example, there are short-term associations of the contributions of indoor sources with
7 ambient concentrations. Recent analysis of PTEAM data (Mage et al., 1999; Wilson et al., 2000)
8 and theoretical considerations (Ott et al., 2000) indicate that it is unlikely that nonambient
9 exposures will be correlated with the ambient concentration (even though total lung dose will be
10 influenced both by ambient and nonambient PM sources and concentrations). Therefore, this
11 type of bias is unlikely. However, if the community average exposure to ambient PM is less than
12 the ambient concentration, the risk regression coefficient, Pc, will be biased low. According to
13 Carroll et al. (1995), Pc = a PE, where Pc is the percentage increase in risk because of a unit
14 increase in ambient concentration, and PE is the estimated percentage increase in risk because of
15 a unit increase in the community-average personal exposure to ambient PM. Both Zeger et al.
16 (2000) and Dominici et al. (2000) examine the nature of error with this second component. Both
17 of these analyses conclude that the error introduced because of measured differences between the
18 average personal exposure and ambient levels can bias the regression coefficients. In both cases
19 they find the Pc is close to a PE.
20 This framework analysis demonstrates the importance of the daily community-average
21 exposure, Et, in community time-series epidemiology. It is Et, not the random, pooled values of
22 E; t, that need to have a statistically significant correlation with Ct for proper interpretation of
23 community time-series epidemiology studies based on ambient monitoring data, as discussed
24 further in Wilson et al. (2000) and Mage et al. (1999).
25 A critical assumption in the above analysis is that the risk varies linearly with C or E (i.e.,
26 Pc and PE are constant). This assumption does not permit a threshold (a concentration below
27 which there is no effect). It also includes the assumption that the appropriate metric for
28 determination of a health response is the 24-h average PM mass concentration. Zeger et al.
29 (2000) show that the likely consequence of using ambient concentrations instead of the risk-
30 weighted average personal exposure measures is to underestimate the pollution effects.
31 According to Zeger et al. (2000) the largest biases in inferences about the mortality-personal
April 2002 5-98 DRAFT-DO NOT QUOTE OR CITE
-------
1 exposure relative risk will occur because of more complex errors between ambient concentration
2 and daily-average personal exposure measures. It is important to note that both the Zeger et al.
3 (2000) and the Dominici et al. (2000) error analyses used personal PM10 data from the PTEAM
4 study data. However, effects of measurement error estimates may differ by particle size and
5 composition. It is possible that PM25, ultrafine particle measures, or another component of PM,
6 may better reflect personal exposures to PM of outdoor origin. Finally, the seasonal or temporal
7 variations in the measurement errors and correlations between different PM concentration
8 measures and gaseous co-pollutants (e.g. SO2, CO, NO2, O3) could influence the error analysis
9 results reported by the investigators cited above.
10
11 5.6.5.2 Analysis of Exposure Measurement Errors in Long-Term Epidemiology Studies
12 The Six Cities (Dockery et al., 1993) and ACS (Pope et al., 1995) studies have played an
13 important role in assessing the health effects from long-term exposures to particulate pollution.
14 Even though these studies often have been considered as chronic epidemiologic studies, it is not
15 easy to differentiate the role of historic exposures from those of recent exposures on chronic
16 disease mortality. In the Six Cities study, fine particles and sulfates were measured at the
17 community level, and the final analysis of the database used six city-wide average ambient
18 concentration measurements. This limitation also applies to the ACS study but has less impact
19 because of the larger number of cities considered in that study. In a HEI-sponsored reanalysis of
20 the Six Cities and the ACS data sets, Krewski et al. (2000) attempted to examine some of the
21 exposure misclassification issues either analytically or through sensitivity analysis of the
22 aerometric and health data. The HEI reanalysis project also addressed exposure measurement
23 error issues related to the Six Cities study. For example, the inability to account for exposures
24 prior to the enrollment of the cohort hampered accurate interpretation of the relative risk
25 estimates in terms of acute versus chronic causes. Although the results seem to suggest past
26 exposures are more strongly associated with mortality than recent exposures, the measurement
27 error for long-term averages could be higher, thus influencing these interpretations. For example,
28 Krewski et al. (2000), using the individual mobility data available for the Six Cities cohort,
29 analyzed the mover and nonmover groups separately. The relative risk of fine particle effects on
30 all-cause mortality was shown to be higher for the nonmover group than for the mover group,
31 suggesting the possibility of higher exposure misclassification biases for the movers. The issue
April 2002 5-99 DRAFT-DO NOT QUOTE OR CITE
-------
1 of using selected ambient monitors in the epidemiologic analyses also was investigated by the
2 ACS and Six Cities studies reanalysis team. Krewski et al.(2000) presented the sensitivity of
3 results to choices made in selecting stationary or mobile-source-oriented monitors. For the ACS
4 study, reanalysis of the sulfate data using only those monitors designated as residential or urban,
5 and excluding sites designated as industrial, agricultural, or mobile did not change the risk
6 estimates appreciably. On the other hand, application of spatial analytic methods designed to
7 control confounding at larger geographic scales (i.e., between cities) caused changes in the
8 particle and sulfate risk coefficients. Spatial adjustment may account for differences in pollution
9 mix or PM composition, but many other cohort-dependent risk factors will vary across regions or
10 cities in the United States. Therefore, it is difficult to interpret these findings solely in terms of
11 spatial differences in pollution composition or relative PM toxicity until further research is
12 concluded.
13 Another study that has examined the influence of measurement errors in air pollution
14 exposure and health effects assessments is the one reported by Navidi et al. (1999). This study
15 developed techniques to incorporate exposure measurement errors encountered in long-term air
16 pollution health effects studies and tested them on the data from the University of Southern
17 California Children's Health Study conducted in 12 communities in California. These
18 investigators developed separate error analysis models for direct (i.e., personal sampling) and
19 indirect (i.e., microenvironmental) personal exposure assessment methods. These models were
20 generic to most air pollutants, but a specific application was performed using a simulated data set
21 for studying ozone health effects on lung function decline in children. Because the assumptions
22 made in their microenvironmental simulation modeling framework were similar to those made in
23 estimating personal PM exposures, it is useful to consider the conclusions from Navidi et. al.
24 (1999). According to Navidi et al. (1999), neither the microenvironmental nor the personal
25 sampler method produces reliable estimates of the exposure-response slope (for O3) when
26 measurement error is uncorrected. Because of nondifferential measurement error, the bias was
27 toward zero under the assumptions made in Navidi et al. (1999) but could be away from zero if
28 the measurement error was correlated with the health response. A simulation analysis indicated
29 that the standard error of the estimate of a health effect increases as the errors in exposure
30 assessment increase (Navidi et al., 1999). According to Navidi et al. (1999), when a fraction of
31 the ambient level in a microenvironment is estimated with a standard error of 30%, the standard
April 2002 5-100 DRAFT-DO NOT QUOTE OR CITE
-------
1 error of the estimate is 50% higher than it would be if the true exposures were known. It appears
2 that errors in estimating ambient PM indoor/ambient PM outdoor ratios have much more
3 influence on the accuracy of the microenvironmental approach than do errors in estimating time
4 spent in these microenvironments.
5
6 5.6.5.3 Conclusions from Analysis of Exposure Measurement Errors on Particulate Matter
7 Epidemiology
8 Personal exposures to PM are influenced by a number of factors and sources of PM located
9 in both indoor and outdoor microenvironments. However, PM resulting from ambient sources
10 does penetrate into indoor environments, such as residences, offices, public buildings, etc., in
11 which individuals spend a large portion of their daily lives. The correlations between total
12 personal exposures and ambient or outdoor PM concentrations can vary depending on the relative
13 contributions of indoor PM sources to total personal exposures. Panel studies of both adult and
14 young subjects have shown that, in fact, individual correlations of personal exposures with
15 ambient PM concentrations could vary person to person, and even day to day, depending on the
16 specific activities of each person. Separation of PM exposures into two components,
17 ambient-generated PM and nonambient PM, would reduce uncertainties in the analysis and
18 interpretation of PM health effects data. Nevertheless, because ambient-generated PM is an
19 integral component of total personal exposures to PM, statistical analysis of cohort-average
20 exposures are strongly correlated with ambient PM concentrations when the size of the
21 underlying population studied is large. Using the PTEAM study data, analysis of exposure
22 measurement errors, in the context of time-series epidemiology, also has shown that errors or
23 uncertainties introduced by using surrogate exposure variables, such as ambient PM
24 concentrations, could lead to biases in the estimation of health risk coefficients. These then
25 would need to be corrected by suitable calibration of the PM health risk coefficients.
26 Correlations between the PM exposure variables and other covariates (e.g., gaseous
27 co-pollutants, weather variables, etc.) also could influence the degree of bias in the estimated PM
28 regression coefficients. However, most time-series regression models employ seasonal or
29 temporal detrending of the variables, thus reducing the magnitude of this cross-correlation
30 problem (Ozkaynak and Spengler 1996).
April 2002 5-101 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ordinarily, exposure measurement errors are not expected to influence the interpretation of
2 findings from either the cross-sectional or time-series epidemiologic studies that have used
3 ambient concentration data if they include sufficient adjustments for seasonality and key
4 confounders. Clearly, there is no question that better estimates of exposures to components of
5 PM of health concern are beneficial. Composition of PM may vary in different geographic
6 locations and different exposure microenvironments. Compositional and spatial variations could
7 lead to further errors in using ambient PM measures as surrogates for exposures to PM. Even
8 though the spatial variability of PM (PM2 5 in particular) mass concentrations in urban
9 environments seems to be small, the same conclusions drawn above regarding the influence of
10 measurement errors may not necessarily hold for all of the PM toxic components. Again, the
11 expectation based on statistical modeling considerations is that these exposure measurement
12 errors or uncertainties will most likely reduce the statistical power of the PM health effects
13 analysis, making it difficult to detect a true underlying association between the correct exposure
14 metric and the health outcome studied. However, until more data on exposures to toxic agents of
15 PM become available, existing studies on PM exposure measurement errors must be relied on;
16 thus, at this time, the working hypothesis is that the use of ambient PM concentrations as a
17 surrogate for exposures is not expected to change the principal conclusions from PM
18 epidemiologic studies, utilizing community average health and pollution data.
19
20
21 5.7 SUMMARY OF KEY FINDINGS AND LIMITATIONS
22 Exposure Definitions and Components
23 • Personal exposure (E) to PM mass or its constituents results when individuals come in contact
24 with particulate pollutant concentrations (C) in locations or microenvironments (jue) that they
25 frequent during a specific period of time. Various PM exposure metrics can be defined
26 according to its source (i.e., ambient, nonambient) and the microenvironment where exposure
27 occurs.
28 • Personal exposure to PM results from an individual's exposure to PM in many different types
29 of microenvironments (e.g., outdoors near home, outdoors away from home, indoors at home,
30 indoors at office or school, commuting, restaurants, malls, other public places, etc.). Thus, total
April 2002 5-102 DRAFT-DO NOT QUOTE OR CITE
-------
1 daily exposure to PM for a single individual (Et) can be expressed as the sum of various
2 microenvironmental exposures that the person encounters during the course of a day.
3 • In a given (j,e, particles may originate from a wide variety of sources. In an indoor
4 microenvironment, PM may be generated from within as a result of PM generating activities
5 (e.g., cooking, cleaning, smoking, resuspending PM from PM resulting from both indoor and
6 outdoor sources that had settled out), from outside (outdoor PM entering through cracks and
7 openings in the structure), and from the chemical interaction of pollutants from outdoor air with
8 indoor-generated pollutants.
9 • The total daily exposure to PM for a single individual (Et) also can be expressed as the sum of
10 contributions of ambient-generated (Eag) and nonambient-generated (Enonag) PM (i.e.,
11 E = Eag+ Enonag). Enonag, in turn, is composed of PM generated by indoor sources (Epig ) and PM
12 generated by personal activities (Epact) (i.e., Enonag = Epig + Epact). Eag is composed of exposures
13 to ambient PM concentrations while outdoors,^ Cakta, and ambient PM that has infiltrated
t
14 indoors,^ Ca;A^ while indoors (i.e., Eag = ^ Cahta + ^ Caihtt). However, within a large
t t t
15 population group, there will be distributions of Et and its components (Eag, Enonag) due to
16 variations in human activities and microenvironmental concentrations and sources each
17 individual encounters.
18 • Exposure models are useful tools for examining the importance of sources, microenvironments,
19 and physical and behavioral factors that influence personal exposures to PM. However,
20 development and evaluation of population exposure models for PM and its components has
21 been limited. Improved modeling methodologies and new model input data are needed.
22
23 Factors Affecting Concentrations and Exposures to Particulate Matter
24 • Concentrations of PM indoors are affected by several factors and mechanisms: ambient
25 concentrations outdoors; air exchange rates; particle penetration factors; particle production
26 from indoor sources and indoor air chemistry; and indoor particle decay rates and removal
27 mechanisms caused by physical processes or resulting from mechanical filtration, ventilation or
28 air-conditioning devices.
29 • Average personal exposures to PM mass and its constituents are influenced by
30 microenvironmental PM concentrations and by how much time is spent by each individual in
April 2002 5-103 DRAFT-DO NOT QUOTE OR CITE
-------
1 these various indoor and outdoor microenvironments. Nationwide, individuals, on average,
2 spend nearly 90% of their time indoors (at home and in other indoor locations) and about 6% of
3 their time outdoors.
4 • Personal exposures are associated with both indoor as well as outdoor sources; the personal
5 exposure/outdoor concentration ratios present substantial intra- and inter-personal variability;
6 although this variability was originally thought to be mainly due to the presence of personal
7 and microenvironmental sources, the results from recent exposure studies suggest that it is the
8 varying impact of the outdoor particles on indoor environments that is mainly responsible for
9 the observed intra- and inter-variability in exposure/outdoor concentration ratios
10 • Home characteristics may be the most important factor that effects the relationship between the
11 average population exposures and ambient concentrations. Air exchange rate seems to be an
12 important home characteristic surrogate that can explain a large fraction of the observe inter-
13 and intra-personal variability. These findings explain why longitudinal studies (many repeated
14 measurements per person) provide stronger correlations between personal exposure and
15 outdoor concentrations than cross-sectional studies (few repeated measurements per
16 individual).
17 • Since home characteristics is the most important factor affecting personal exposures, one
18 would expect that correlations between average population exposures and outdoor
19 concentrations will vary by season and geography.
20 • The relative size of personal exposure to ambient-generated PM relative to nonambient-
21 generated PM depends on the ambient concentration, the infiltration rate of outdoor PM into
22 indoor microenvironments, the amount of PM generated indoors (e.g., ETS, cooking and
23 cleaning emissions), and the amount of PM generated by personal activity sources. Infiltration
24 rates primarily depend on air-exchange rate, size-dependent particle penetration across the
25 building membrane, and size-dependent removal rates. All of these factors vary over time and
26 across subjects and building types.
27 • The relationship between PM exposure, dose, and health outcome could depend on the
28 concentration, composition, and toxicity of the PM originating from different sources.
29 Application of source apportionment techniques to indoor and outdoor PM2 5 and personal,
30 indoor, and outdoor PM10 composition data have identified the following general source
April 2002 5-104 DRAFT-DO NOT QUOTE OR CITE
-------
1 categories: outside soil, resuspended indoor soil, indoor soil, personal activities, sea-salt,
2 motor vehicles, nonferrous metal smelters, and secondary sulfates.
3 • There have been only a limited number of studies that have measured the physical and
4 chemical constituents of PM in personal or microenvironmental samples. Available data on
5 PM constituents indicate that
6 S personal and indoor sulfate measurements often are correlated highly with outdoor and
7 ambient sulfate concentration measurements;
8 S for acid aerosols, indoor air chemistry is particularly important because of the
9 neutralization of the acidity by ammonia, which is present at higher concentrations
10 indoors because of the presence of indoor sources of ammonia;
11 S for SVOCs, including PAHs and phthalates, the presence of indoor sources will
12 substantially impact the relation between indoor and ambient concentrations;
13 S penetration and decay rates are a functions of size and will cause variations in the
14 attenuation factors as a function of particle size; infiltration rates will be higher for PMX
15 and PM2 5 than for PM10, PM10_2 5 or ultrafme particles; and
16 S Indoor air chemistry may increase indoor concentrations of organic PM.
17 • Even though there is an increasing amount of research being performed to measure PM
18 constituents in different PM size fractions, with few exceptions (i.e., sulfur or sulfates), the
19 current data are inadequate to adequately assess the relationship between personal, indoor, and
20 ambient concentrations of most PM constituents.
21
22 Correlations Between Personal Exposures, Indoor, Outdoor, and Ambient Measurements
23 • Most of the available personal data on PM measurements and information on the relationships
24 between personal and ambient PM come from a few large-scale studies, such as the PTEAM
25 study, or the longitudinal panel studies, which have been conducted on selected populations,
26 such as the elderly.
27 • Panel and cohort studies that have measured PM exposures and concentrations typically have
28 reported their results in terms of three types of correlations: (1) longitudinal, (2) pooled, and
29 (3) daily-average correlations between personal and ambient or outdoor PM.
30 • The type of correlation analysis performed can have a substantial effect on the resulting
31 correlation coefficient. Low correlations with ambient concentrations could result when people
April 2002 5-105 DRAFT-DO NOT QUOTE OR CITE
-------
1 with very different nonambient exposures are pooled, even though temporally, their individual
2 personal exposures may be correlated highly with ambient concentrations.
3 • Recent studies conducted by EPA of the elderly subjects living in a retirement facility in
4 Baltimore and a group of elderly living in Fresno produced higher correlation coefficients
5 between personal and ambient PM for daily-average correlations compared to longitudinal
6 correlations. This supports earlier analyses showing the daily-average correlations are higher
7 than pooled correlations.
8 • Longitudinal and pooled correlations between personal exposure and ambient or outdoor PM
9 concentrations reported by various investigators varied considerably among the different
10 studies and in each study between the study subjects. Most studies report longitudinal
11 correlation coefficients that range from close to zero to near one, indicating that individual's
12 activities and residence type may have a significant effect on total personal exposures to PM.
13 • Longitudinal studies that measured sulfate found high correlations between personal and
14 ambient sulfate.
15 • In general, probability-based population studies tend to show low pooled correlations because
16 of the high differences in levels of nonambient PM generating activities from one subject to
17 another. In contrast, the absence of indoor sources for the populations in several of the
18 longitudinal panel studies resulted in high correlations between personal exposure and ambient
19 PM within subjects over time for these populations. But even for these studies, correlations
20 varied by individual depending on their activities and microenvironments that they occupied.
21
22 Potential Sources of Error Resulting from Using Ambient Particulate Matter
23 Concentrations in Epidemiologic Analyses
24 • There is, as yet, no clear consensus among exposure analysts as to how well ambiently
25 measured PM concentrations represent a surrogate for personal exposure to total PM or to
26 ambient-generated PM.
27 • Measurement studies of personal exposures to PM are still few and limited in spatial, temporal,
28 and demographic coverage. Consequently, with the exception of a few longitudinal panel
29 studies, most epidemiologic studies on PM health effects have relied on daily-average PM
30 concentration measurements obtained from ambient community monitoring data as a surrogate
31 for the exposure variable.
April 2002 5-106 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Because individuals are exposed to particles in a multitude of indoor and outdoor
2 microenvironments during the course of a day, concerns about error introduced in the
3 estimation of PM risk coefficients using ambient, as opposed to personal PM measurements,
4 have been raised.
5 • Total personal exposures to PM could vary from person to person, and even day to day,
6 depending on the specific activities of each person. Separation of PM exposures into two
7 components, ambient-generated PM and nonambient-generated PM, would reduce potential
8 uncertainties in the analysis and interpretation of PM health effects data.
9 • Available data indicate that PM mass concentrations, especially fine PM, typically are
10 distributed uniformly in most metropolitan areas, thus reducing the potential for exposure
11 misclassification because of spatial variability when a limited number of ambient PM monitors
12 are used to represent population average ambient exposures in community time-series or
13 long-term, cross-sectional epidemiologic studies of PM.
14 • Even though the spatial variability of PM (in particular, PM25) mass concentrations in urban
15 environments seems to be small, the same conclusions drawn above regarding the influence of
16 measurement errors may not necessarily hold for all of the PM components.
17 • There are important differences in the relationship of ambient PM concentrations (Ca) with
18 exposures to ambient PM (Eag), and with exposures to nonambient PM (Enonag). Various
19 researchers have shown that Eag is a function of Ca, and that concentrations of ambient PM are
20 driven by meteorology, by changes in source emission rates, and in locations of emission
21 sources relative to the measurement site. However, Enonag is independent of Ca, because
22 concentrations of nonambient PM are driven by the daily activities of people.
23 • Because personal exposures also include a contribution from ambient concentrations, the
24 correlation between daily-average personal exposure and the daily-average ambient
25 concentration increases as the number of subjects measured daily increases. An application of
26 a Random Component Superposition (RCS) model has shown that the contributions of ambient
27 PM10 and indoor-generated PM10 to community mean exposure can be decoupled in modeling
28 urban population exposure distributions.
29 • If linear nonthreshold models are assumed in time-series analysis of daily-average ambient PM
30 concentrations and community health data, Enonag is not expected to contribute to the relative
31 risk estimates determined by regression of health responses on Ca.
April 2002 5-107 DRAFT-DO NOT QUOTE OR CITE
-------
1 • Using the PTEAM study data, analysis of exposure measurement errors in the context of
2 time-series epidemiology has shown that errors or uncertainties introduced by using surrogate
3 exposure variables, such as ambient PM concentrations, could lead to biases in the estimation
4 of health risk coefficients.
5 • Because sources and chemical composition of particulate matter affecting personal exposures in
6 different microenvironments vary, by season, day-of-the-week, and time of day, it is likely that
7 some degree of misclassification of exposures to PM toxic agents of concern will be introduced
8 when health effects models use only daily-average mass measures such as PM10 or PM2 5.
9 Because of the paucity of currently available data on many of these factors, it is impossible to
10 ascertain at this point the significance of these more complex exposure misclassification
11 problems in the interpretation of results from PM epidemiology.
12 • Exposure measurement errors may depend on particle size and composition. PM2 5 better
13 reflects personal exposure to PM of outdoor origin than PM10. It is possible that various
14 ultrafine particle measures, or other components of PM may be better exposure indicators for
15 epidemiologic studies.
16 • Seasonal or temporal variations in the measurement errors and their correlations between
17 different PM concentration measures and co-pollutants (e.g., SO2, CO, NO2, O3) could
18 influence the error analysis results but not likely the interpretation of current findings.
19 • Multi-pollutant personal exposure studies have suggested that ambient concentrations of
20 gaseous copollutants serve as surrogates of personal exposures to particles rather than as
21 confounders.
22 • Ordinarily, PM exposure measurement errors are not expected to influence the interpretation of
23 findings from either the community time-series or long-term epidemiologic studies that have
24 used ambient concentration data if they include sufficient adjustments for seasonality and key
25 personal and geographic confounders.
26 • In the context of long-term epidemiologic studies, it appears that the errors introduced in
27 estimating ambient PM indoor/ambient PM outdoor ratios have much more influence on the
28 accuracy of the microenvironmental exposure estimation approach than do errors in estimating
29 time spent in these microenvironments.
April 2002 5-108 DRAFT-DO NOT QUOTE OR CITE
-------
1 • To reduce exposure misclassification errors in PM epidemiology, conducting new cohort
2 studies of sensitive populations with better real-time techniques for exposure monitoring and
3 further speciation of indoor-generated, ambient, and personal PM mass are essential.
4 • Based on statistical modeling considerations, it is expected that existing PM exposure
5 measurement errors or uncertainties most likely will reduce the statistical power of the PM
6 health effects analysis, thus making it difficult to detect a true underlying association between
7 the correct exposure metric and the health outcome studied.
8 • Although exposure measurement errors for fine particles are not expected to influence the
9 interpretation of findings from either the community time-series or the long-term, cross-
10 sectional epidemiologic studies that have used ambient concentration data, they may
11 underestimate the strength of the impact. Sufficient data are not available to evaluate the
12 impact of exposure measurement error for other PM species or size fractions.
April 2002 5-109 DRAFT-DO NOT QUOTE OR CITE
-------
1 REFERENCES
2 Abt, E.; Suh, H. H.; Allen, G.; Koutrakis, P. (2000a) Characterization of indoor particle sources: a study conducted
3 in the metropolitan Boston area. Environ. Health Perspect. 108: 35-44.
4 Abt, E.; Suh, H. H.; Catalano, P.; Koutrakis, P. (2000b) The relative contribution of outdoor and indoor particle
5 sources to indoor concentrations. Environ. Sci. Technol. 34: 3579-3587.
6 Akland, G. G.; Hartwell, T. D.; Johnson, T. R.; Whitmore, R. W. (1985) Measuring human exposure to carbon
7 monoxide in Washington, D.C., and Denver, Colorado, during the winter of 1982-1983. Environ. Sci.
8 Technol. 19:911-918.
9 Al-Raheem, M. Y. A.; El-Desouky, M; Mage, D. T.; Kollander, M.; Wallace, L. A.; Spengler, J. D. (2000)
10 Personal exposures to PM and gaseous air pollutants at the end of the 1991 Kuwait oil fires. Presented at:
11 PM2000: paniculate matter and health—the scientific basis for regulatory decision-making, specialty
12 conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management Association.
13 Aim, S.; Jantunen, M. J.; Vartiainen, M. (1999) Urban commuter exposure to particle matter and carbon monoxide
14 inside an automobile. J. Exposure Anal. Environ. Epidemiol. 9: 237-244.
15 Anuszewski, J.; Larson, T. V.; Koenig, J. Q. (1998) Simultaneous indoor and outdoor particle light-scattering
16 measurements at nine homes using a portable nephelometer. J. Exposure Anal. Environ. Epidemiol.
17 8:483-493.
18 Baek, S.-O.; Kim, Y.-S.; Perry, R. (1997) Indoor air quality in homes, offices and restaurants in Korean urban
19 areas—indoor/outdoor relationships. Atmos. Environ. 31: 529-544.
20 Bahadori, T.; Suh, H. H.; Rojas, L.; Koutrakis, P. (1999) Characterization of personal PM10 and PM2.5 exposures
21 for individuals with COPD: results from a pilot study in Nashville, TN. J. Exposure Anal. Environ.
22 Epidemiol.: in press.
23 Bahadori, T.; Suh, H. H.; Rojas-Bracho, L.; Koutrakis, P. (2001) Personal exposure to paniculate matter of
24 individuals with chronic obstructive pulmonary disease (COPD). J. Exposure Anal. Environ. Epidemiol.:
25 accepted.
26 Bohadana, A. B.; Massin, N.; Wild, P.; Toamain, J.-P.; Engel, S.; Goutet, P. (2000) Symptoms, airway
27 responsiveness, and exposure to dust in beech and oak wood workers. Occup. Environ. Med. 57: 268-273.
28 Brauer, M.; Mannetje, A. (1998) Restaurant smoking restrictions and environmental tobacco smoke exposure. Am.
29 J. Public Health 88: 1834-1836.
30 Brauer, M.; Koutrakis, P.; Spengler, J. D. (1989) Personal exposures to acidic aerosols and gases. Environ. Sci.
31 Technol. 23: 1408-1412.
32 Brauer, M.; Koutrakis, P.; Keeler, G. J.; Spengler, J. D. (1990) Indoor and outdoor concentrations of acidic aerosols
33 and gases. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and climate,
34 volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
3 5 International Conference on Indoor Air Quality and Climate, Inc.; pp. 447-452.
36 Brauer, M.; Bartlett, K.; Regalado-Pineda, J.; Perez-Padilla, R. (1996) Assessment of paniculate concentrations
37 from domestic biomass combustion in rural Mexico. Environ. Sci. Technol. 30: 104-109.
38 Brauer, M.; Hirtle, R. D.; Hall, A. C.; Yip, T. R. (1999) Monitoring personal fine particle exposure with a particle
39 counter. J. Exposure Anal. Environ. Epidemiol. 9: 228-236.
40 Brauer, M.; Hruba, F.; Mihalikova, E.; Fabianova, E.; Miskovic, P.; Plzikova, A.; Lendacka, M.; Vanderberg, J.;
41 Cullen, A. (2000) Personal exposure to particles in Banska Bystrica, Slovakia. J. Exposure Anal. Environ.
42 Epidemiol. 10: 478-487.
43 Brown, K. G.; Paxton, M. B. (1998) Predictability of personal exposure to airborne paniculate matter from ambient
44 concentrations in four communities. In: Vostal, J. J., ed. Health effects of paniculate matter in ambient air.
45 Proceedings of an international conference; 1997; Prague, Czech Republic. Pittsburgh, PA: Air & Waste
46 Management Association; pp. 333-350. (A&WMA publication VIP-80).
47 Burke, J.; Zufall, M.; Ozkaynak, H. (2001) A population exposure model for paniculate matter: case study results
48 for PM2 5 in Philadelphia, PA. J. Exposure Anal. Environ. Epidemiol. 11: 470-489.
49 Burton, R. M.; Suh, H. H.; Koutrakis, P. (1996) Spatial variation in paniculate concentrations within metropolitan
50 Philadelphia. Environ. Sci. Technol. 30: 400-407.
51 Buzorius, G.; Ha'meri, K.; Pekkanen, J.; Kulmala, M. (1999) Spatial variation of aerosol number concentration in
52 Helsinki city. Atmos. Environ. 33: 553-565.
53 Calder, K. L. (1957) A numerical analysis of the protection afforded by buildings against BW aerosol attack.
54 Fort Detrick, MD: Office of the Deputy Commander for Scientific Activities; BWL tech. study no. 2.
April 2002 5-110 DRAFT-DO NOT QUOTE OR CITE
-------
1 Carrer, P.; Alcini, D.; Cavallo, D.; Bollini, D.; Ghalandar, R.; Lovato, L.; Ploia, E.; Vercelli, F.; Visigalli, F.;
2 Maroni, M. (1998) Personal exposure to PM10 of subjects living in Milan. In: Vostal, J. J., ed. Health effects
3 of paniculate matter in ambient air. Proceedings of an international conference; 1997; Prague, Czech
4 Republic. Pittsburgh, PA: Air & Waste Management Association; pp. 359-365. (A&WMA publication
5 VIP-80).
6 Carroll, R. J.; Ruppert, D.; Stefanski, L. A. (1995) Measurement error in nonlinear models. London, United
7 Kingdom: Chapman & Hall. (Cox, D. R.; Hinkley, D. V.; Keiding, N.; Reid, N.; Rubin, D. B.; Silverman,
8 B. W., eds. Monographs on statistics and applied probability: v. 63).
9 Cha, Q.; Chen, Y.-Z.; Du, Y.-X. (1996) Influence of indoor air pollution on quality of outdoor air in city of
10 Guangzhou, China. In: Indoor air '96: the 7th international conference on indoor air quality, volume I; July;
11 Nagoya, Japan. Tokyo, Japan: Institute of Public Health; pp. 1131-1135.
12 Chang, L.-T.; Koutrakis, P.; Catalano, P. J.; Suh, H. H. (2000) Hourly personal exposures to fine particles and
13 gaseous pollutants - results from Baltimore, Maryland. J. Air Waste Manage. Assoc. 50: 1223-1235.
14 Chao, C. Y. H.; Tung, T. C. (2001) An empirical model for outdoor contaminant transmission into residential
15 buildings and experimental verification. Atmos. Environ. 35: 1585-1596.
16 Chuang, J. C.; Callahan, P. J.; Lyu, C. W.; Wilson, N. K. (1999) Polycyclic aromatic hydrocarbon exposures of
17 children in low-income families. J. Exposure Anal. Environ. Epidemiol. 2: 85-98.
18 Clayton, C. A.; Perritt, R. L.; Pellizzari, E. D.; Thomas, K. W.; Whitmore, R. W.; Wallace, L. A.; Ozkaynak, H.;
19 Spengler, J. D. (1993) Particle total exposure assessment methodology (PTEAM) study: distributions of
20 aerosol and elemental concentrations in personal, indoor, and outdoor air samples in a southern California
21 community. J. Exposure Anal. Environ. Epidemiol. 3: 227-250.
22 Clayton, C. A.; Pellizzari, E. D.; Rodes, C. E.; Mason, R. E.; Piper, L. L. (1999a) Estimating distributions of
23 long-term paniculate matter and manganese exposures for residents of Toronto, Canada. Atmos. Environ.
24 33:2515-2526.
25 Clayton, C. A.; Pellizzari, E. D.; Whitmore, R. W.; Perritt, R. L., Quackenboss, J. J. (1999b) National human
26 exposure assessment survey (NHEXAS): distributions and associations of lead, arsenic, and volatile organic
27 compounds in EPA Region 5. J. Exposure Anal. Environ. Epidemiol. 9: 381-392.
28 Cohen, B. S.; Li, W.; Xiong, J. Q.; Lippmann, M. (2000) Detecting H+ in ultrafine ambient aerosol using iron
29 nano-film detectors and scanning probe microscopy. Appl. Occup. Environ. Hyg. 15: 80-89.
30 Conner, T. L.; Lonneman, W. A.; Seila, R. L. (1995) Transportation-related volatile hydrocarbon source profiles
31 measured in Atlanta. J. Air Waste Manage. Assoc. 45: 383-394.
32 Copeland, K. T.; Checkoway, H.; McMichael, A. J.; Holbrook, R. H. (1977) Bias due to misclassification in the
33 estimation of relative risk. Am. J. Epidemiol. 105: 488-495.
34 Crump, K. S. (2000) Manganese exposures in Toronto during use of the gasoline additive, methylcyclopentadienyl
35 manganese tricarbonyl. J. Exposure Anal. Environ. Epidemiol. 10: 227-239.
36 Daisey, J. M.; Mahanama, K. R. R.; Hodgson, A. T. (1998) Toxic volatile organic compounds in simulated
37 enviromental tobacco smoke: emission factors for exposure assessment. J. Exposure Anal. Environ.
38 Epidemiol. 8:313-334.
39 Dockery, D. W.; Spengler, J. D. (1981) Indoor-outdoor relationships of respirable sulfates and particles.
40 Atmos. Environ. 15: 335-343.
41 Dockery, D. W.; Pope, C. A., Ill; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
42 (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med.
43 329: 1753-1759.
44 Dominici, F.; Zeger, S. L.; Samet, J. (2000) A measurement error model for time-series studies of air pollution and
45 mortality. Biostatistics 1: 157-175.
46 Donham, K. J.; Cumro, D.; Reynolds, S. J.; Merchant, J. A. (2000) Dose-response relationships between
47 occupational aerosol exposures and cross-shift declines of lung function in poultry workers:
48 recommendations for exposure limits. J. Occup. Environ. Med. 42: 260-269.
49 Duan, N. (1982) Models for human exposure to air pollution. Environ. Int. 8: 305-309.
50 Duan, N. (1991) Stochastic microenvironment models for air pollution exposure. J. Exposure Anal. Environ.
51 Epidemiol. 1:235-257.
52 Dubowsky, S. D.; Wallace, L. A.; Buckley, T. J. (1999) The contribution of traffic to indoor concentrations of
53 polycyclic aromatic hydrocarbons. J. Exposure Anal. Environ. Epidemiol. 9: 312-321.
April 2002 5-111 DRAFT-DO NOT QUOTE OR CITE
-------
1 Ebelt, S. T.; Petkau, A. J.; Vedal, S.; Fisher, T. V.; Brauer, M. (2000) Exposure of chronic obstructive pulmonary
2 disease patients to paniculate matter: relationships between personal and ambient air concentrations. J. Air
3 Waste Manage. Assoc. 50: 1081-1094.
4 Esmen, N. A.; Hall, T. A. (2000) Theoretical investigation of the interrelationship between stationary and personal
5 sampling in exposure estimation. Appl. Occup. Environ. Hyg. 15: 114-119.
6 Evans, G. F.; Highsmith, R. V.; Sheldon, L. S.; Suggs, J. C.; Williams, R. W.; Zweidinger, R. B.; Creason, J. P.;
7 Walsh, D.; Rodes, C. E.; Lawless, P. A. (2000) The 1999 Fresno paniculate matter exposure studies:
8 comparison of community, outdoor, and residential PM mass measurements. J. Air Waste Manage. Assoc.
9 50: 1887-1896.
10 Ezzati, M.; Kammen, D. M. (2001) Quantifying the effects of exposure to indoor air pollution from biomass
11 combustion on acute respiratory infections in developing countries. Environ. Health Perspect. 109: 481-488.
12 Ferris, B. G., Jr.; Speizer, F. E.; Spengler, J. D.; Dockery, D.; Bishop, Y. M. M.; Wolfson, M.; Humble, C. (1979)
13 Effects of sulfur oxides and respirable particles on human health: methodology and demography of
14 populations in study. Am. Rev. Respir. Dis. 120: 767-779.
15 Fischer, P. H.; Hoek, G.; Van Reeuwijk, H.; Briggs, D. J.; Lebret, E.; Van Wijnen, J. H.; Kingham, S.; Elliott, P. E.
16 (2000) Traffic-related differences in outdoor and indoor concentrations of particles and volatile organic
17 compounds in Amsterdam. Atmos. Environ. 34: 3713-3722.
18 Freijer, J. L; Bloemen, H. J. T. (2000) Modeling relationships between indoor and outdoor air quality. J. Air Waste
19 Manage. Assoc. 50: 292-300.
20 Frey, H. C.; Rhodes, D. S. (1996) Characterizing, simulating and analyzing variability and uncertainty:
21 an illustration of methods using an air toxic emissions example. Hum. Ecol. Risk Assess. 2: 762-797'.
22 Gladen, B.; Rogan, W. J. (1979) Misclassification and the design of environmental studies. Am. J. Epidemiol.
23 109:607-616.
24 Glen, G.; Lakkadi, Y.; Tippett, J. A.; del Valle-Torres, M. (1997) Development of NERL/CHAD: the National
25 Exposure Research Laboratory consolidated human activity database. Research Triangle Park, NC: U.S.
26 Environmental Protection Agency, Office of Research and Development; contract no. 68-D5-0049.
27 Goldstein, B. R.; Tardiff, G.; Hoffnagle, G.; Kester, R. (1992) Valdez health study: summary report. Ankorage, AK:
28 Alyeska Pipeline Service Company.
29 Hadwen, G. E.; McCarthy, J. F.; Womble, S. E.; Girman, J. R.; Brightman, H. S. (1997) Volatile organic compound
30 concentrations in 41 office buildings in the continental United States. In: Woods, J. E.; Grimsrud, D. T.;
31 Boschi, N., eds. Proceedings of healthy buildings/indoor air quality international conference; September;
32 Bethesda, MD. Washington, DC: Healthy Buildings/IAQ. [Healthy Buildings - International Conference;
33 v. 2, pp. 465-470.]
34 Hartwell, T. D.; Clayton, C. A.; Michie, R. M., Jr.; Whitmore, R. W.; Zelon, H. S.; Whitehurst, D. A.; Akland,
35 G. G. (1984) Study of carbon monoxide exposures of residents of Washington, D.C. Research Triangle Park,
36 NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
37 EPA-600/D-84-177. Available from: NTIS, Springfield, VA; PB84-230051.
38 Hayes, S. R.; Marshall, J. R. (1999) Designing optimal strategies to attain the new U.S. paniculate matter standards:
39 some initial concepts. J. Air Waste Manage. Assoc. 49: PM192-PM198.
40 Henry, R. C.; Lewis, C. W.; Collins, J. F. (1994) Vehicle-related source compositions from ambient data:
41 the GRACE/SAFER method. Environ. Sci. Technol. 28: 823-832.
42 Houseman, E. A.; Ryan, L.; Levy, J.; Richardson, D.; Spenger, J. D. (2001) Autocorrelation in real-time continuous
43 monitoring of microenvironments. J. Appl. Stat: in press.
44 Howard-Reed, C.; Rea, A. W.; Zufall, M. J.; Burke, J. M.; Williams, R. W.; Suggs, J. C.; Sheldon, L. S.; Walsh, D.;
45 Kwock, R. (2000) Use of a continuous nephelometer to measure personal exposure to particles during the
46 U.S. Environmental Protection Agency Baltimore and Fresno panel studies. J. Air Waste Manage. Assoc.
47 50: 1125-1132.
48 Institute for Social Research. (1997) Child development supplement to the panel study of income dynamics.
49 Ann Arbor, MI: University of Michigan. Available: www.isr.umich.edu/src/child-development/home.html
50 [2001, January 31].
51 Isukapalli, S. S.; Georgopoulos, P. G. (2000) Modeling of outdoor/indoor relationships of gas phase pollutants and
52 paniculate matter. Presented at: PM2000: paniculate matter and health—the scientific basis for regulatory
53 decision-making, specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste
54 Management Association.
April 2002 5-112 DRAFT-DO NOT QUOTE OR CITE
-------
1 Jang, M; Kamens, R. M. (1999) Newly characterized products and composition of secondary aerosols from the
2 reaction of a-pinene with ozone. Atmos. Environ. 33: 459-474.
3 Janssen, N. (1998) Personal exposure to airborne particles: validity of outdoor concentrations as a measure of
4 exposure in time series studies [dissertation]. Wageningen, Netherlands: Landbouwuniversiteit Wageningen.
5 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1995) A relation between personal and ambient.
6 Epidemiology 6(suppl.): S45.
7 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1997) Childhood exposure to PM10: relation between
8 personal, classroom, and outdoor concentrations. Occup. Environ. Med. 54: 888-894.
9 Janssen, N. A. H.; Hoek, G.; Brunekreef, B.; Harssema, H.; Mensink, I.; Zuidhof, A. (1998a) Personal sampling of
10 particles in adults: relation among personal, indoor, and outdoor air concentrations. Am. J. Epidemiol.
11 147:537-547.
12 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1998b) Personal sampling of airborne particles:
13 method performance and data quality. J. Exposure Anal. Environ. Epidemiol. 8: 37-49.
14 Janssen, N. A. H.; Hoek, G.; Harssema, H.; Brunekreef, B. (1999a) Personal exposure to fine particles in children
15 correlates closely with ambient fine particles. Arch. Environ. Health 54: 95-101.
16 Janssen, N. A. H.; Hoek, G.; Brunekreef, B.; Harssema, H. (1999b) Mass concentration and elemental composition
17 of PM10 in classrooms. Occup. Environ. Med. 56: 482-487.
18 Janssen, L. H. J. M.; Buringh, E.; van der Meulen, A.; van den Hout, K. D. (1999c) A method to estimate the
19 distribution of various fractions of PM10 in ambient air in the Netherlands. Atmos. Environ. 33: 3325-3334.
20 Janssen, N. A. H.; De Hartog, J. J.; Hoek, G.; Brunekreef, B.; Lanki, T.; Timonen, K. L.; Pekkanen, J. (2000)
21 Personal exposure to fine paniculate matter in elderly subjects: relation between personal, indoor, and
22 outdoor concentrations. J. Air Waste Manage. Assoc. 50: 1133-1143.
23 Jantunen, M. J.; Hanninen, O.; Katsouyanni, K.; Knoppel, H.; Kuenzli, N.; Lebret, E.; Maroni, M.; Saarela, K.;
24 Sram, R.; Zmirou, D. (1998) Air pollution exposure in European cities: the EXPOLIS study. J. Exposure
25 Anal. Environ. Epidemiol. 8: 495-518.
26 Jantunen, M. J.; Kousa, A.; Hanninen, O.; Koistinen, K.; Sram, R.; Lanki, T.; Lioy, P. J. (2000) Long term
27 population PM exposures correlate well with outdoor air levels - short term individual exposures do not.
28 Presented at: PM2000: paniculate matter and health—the scientific basis for regulatory decision-making,
29 specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA: Air & Waste Management
30 Association.
31 Jenkins, R. A.; Palausky, M. A.; Counts, R. W.; Guerin, M. R.; Dindal, A. B.; Bayne, C. K. (1996a) Determination
32 of personal exposure of non-smokers to environmental tobacco smoke in the United States. Lung Cancer
33 14(suppl. 1): S195-S213.
34 Jenkins, R. A.; Palausky, A.; Counts, R. W.; Bayne, C. K.; Dindal, A. B.; Guerin, M. R. (1996b) Exposure to
3 5 environmental tobacco smoke in sixteen cities in the United States as determined by personal breathing zone
36 air sampling. J. Exposure Anal. Environ. Epidemiol. 6: 473-502.
37 Johnson, T. (1984) A study of personal exposure to carbon monoxide in Denver, Colorado. Research Triangle Park,
3 8 NC: U.S. Environmental Protection Agency, Environmental Monitoring Systems Laboratory; report no.
39 EPA-600/4-84-014.
40 Johnson, T. (1989) Human activity patterns in Cincinnati, Ohio [final report]. Palo Alto, CA: Electric Power
41 Research Institute; project no. RP940-06; report no. EN-6204.
42 Johnson, T. R. (1995) Recent advances in the estimation of population exposure to mobile source pollutants.
43 J. Exposure Anal. Environ. Epidemiol. 5: 551-571.
44 Johnson, T.; Long, T.; Ollison, W. (2000) Prediction of hourly microenvironmental concentrations of fine particles
45 based on measurements obtained from the Baltimore scripted activity study. J. Exposure Anal. Environ.
46 Epidemiol. 10:403-411.
47 Jones, N. C.; Thornton, C. A.; Mark, D.; Harrison, R. M. (2000) Indoor/outdoor relationships of paniculate matter
48 in domestic homes with roadside, urban and rural locations. Atmos. Environ. 34: 2603-2612.
49 Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
50 segregated mass concentration data and ultrafine particle number concentrations in urban areas.
51 Atmos. Environ. 33: 2907-2913.
52 Kingham, S.; Briggs, D.; Elliott, P.; Fischer, P.; Lebret, E. (2000) Spatial variations in the concentrations of
53 traffic-related pollutants in indoor and outdoor air in Huddersfield, England. Atmos. Environ. 34: 905-916.
54 Kleeman, M. J.; Cass, G. R. (1998) Source contributions to the size and composition distribution of urban
55 paniculate air pollution. Atmos. Environ. 32: 2803-2816.
April 2002 5-113 DRAFT-DO NOT QUOTE OR CITE
-------
1 Klepeis, N. E. (1999) An introduction to the indirect exposure assessment approach: modeling human exposure
2 using microenvironmental measurements and the recent national human activity pattern survey. Environ.
3 Health Perspect. Suppl. 107(2): 365-374.
4 Klepeis, N. E.; Ott, W. R.; Switzer, P. (1994) A Total Human Exposure Model (THEM) for Respirable Suspended
5 Particles (RSP). Presented at: 87th annual meeting & exhibition of the Air & Waste Management
6 Association; June; Cincinnati, OH. Pittsburgh, PA: Air & Waste Management Association; paper no. 94-
7 WA75A.03. Available from: NTIS, Springfield, VA; PB94-197415.
8 Klepeis, N. E.; Ott, W. R.; Switzer, P. (1995) Modeling the time series of respirable suspended particles and carbon
9 monoxide from multiple smokers: validation in two public smoking lounges. Presented at: 88th annual
10 meeting and exhibition of the Air & Waste Management Association; June; San Antonio, TX. Pittsburgh,
11 PA: Air & Waste Management Association; paper no. 95-MPH.03.
12 Klepeis, N. E.; Ott, W. R.; Switzer, P. (1996) A multiple-smoker model for predicting indoor air quality in public
13 lounges. Environ. Sci. Technol. 30: 2813-2820.
14 Klepeis, N. E.; Switzer, P.; Ott, W. R. (2000) Estimating exposure to environmental tobacco smoke particles using
15 24-hour time-activity recall diaries. Presented at: PM2000: paniculate matter and health—the scientific basis
16 for regulatory decision-making, specialty conference & exhibition; January; Charleston, SC. Pittsburgh, PA:
17 Air & Waste Management Association.
18 Koontz, M. D.; Rector, H. E. (1995) Estimation of distributions for residential air exchange rates: final report.
19 Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics;
20 contract nos. 68-D9-0166 and 68-D3-0013.
21 Koontz, M. D.; Niang, L. L. (1998) California population indoor exposure model (CPIEM): version 1.4F. Final
22 report A933-157: user's guide. Sacramento, CA: California Air Resources Board; GEOMET report no.
23 IE-2632.
24 Koutrakis, P.; Briggs, S. L. K.; Leaderer, B. P. (1992) Source apportionment of indoor aerosols in Suffolk and
25 Onondaga Counties, New York. Environ. Sci. Technol. 26: 521-527.
26 Krewski, D.; Burnett, R. T.; Goldberg, M. S.; Hoover, K.; Siemiatycki, J.; Jerrett, M.; Abrahamowicz, M.;
27 White, W. H. (2000) Reanalysis of the Harvard Six Cities study and the American Cancer Society study of
28 paniculate air pollution and mortality: a special report of the Institute's Particle Epidemiology Reanalysis
29 Project. Cambridge, MA: Health Effects Institute.
30 Lachenmyer, C.; Hidy, G. M. (2000) Urban measurements of outdoor-indoor PM2 5 concentrations and personal
31 exposure in the deep south. Part I. Pilot study of mass concentrations for nonsmoking subjects. Aerosol Sci.
32 Technol. 32: 34-51.
33 Laden, F.; Neas, L. M.; Dockery, D. W.; Schwartz, J. (2000) Association of fine paniculate matter from different
34 sources with daily mortality in six U.S. cities. Environ. Health Perspect. 108: 941-947.
35 Leaderer, B. P.; Naeher, L.; Jankun, T.; Balenger, K.; Holford, T. R.; Toth, C.; Sullivan, J.; Wolf son, J. M.;
36 Koutrakis, P. (1999) Indoor, outdoor, and regional summer and winter concentrations of PM10, PM25, SO42",
37 H+, NH4+, NO3", NH3, and nitrous acid in homes with and without kerosene space heaters. Environ. Health
38 Perspect. 107:223-231.
39 Lee, S. C.; Chang, M. (1999) Indoor air quality investigations at five classrooms. Indoor Air 9: 134-138.
40 Lessler, J. T.; Kalsbeek, W. D. (1992) Nonsampling error in surveys. New York, NY: John Wiley & Sons, Inc.;
41 p. 7.
42 Liao, D.; Creason, J.; Shy, C.; Williams, R.; Watts, R.; Zweidinger, R. (1999) Daily variation of paniculate air
43 pollution and poor cardiac autonomic control in the elderly. Environ. Health Perspect. 107: 521-525.
44 Lillquist, D. R.; Lee, J. S.; Ramsay, J. R.; Boucher, K. M.; Walton, Z. L.; Lyon, J. L. (1998) A comparison of
45 indoor/outdoor PM10 concentrations measured at three hospitals and a centrally located monitor in Utah.
46 Appl. Occup. Environ. Hyg. 13: 409-415.
47 Linn, W. S.; Gong, H., Jr.; Clark, K. W.; Anderson, K. R. (1999) Day-to-day paniculate exposures and health
48 changes in Los Angeles area residents with severe lung disease. J. Air Waste Manage. Assoc.
49 49:PM108-PM115.
50 Lioy, P. J. (1990) Assessing total human exposure to contaminants: a multidisciplinary approach. Environ. Sci.
51 Technol. 24: 938-945.
52 Lioy, P. J.; Waldman, J. M.; Buckley, T.; Butler, J.; Pietarinen, C. (1990) The personal, indoor and outdoor
53 concentrations of PM-10 measured in an industrial community during the winter. Atmos. Environ. Part B
54 24:57-66.
April 2002 5-114 DRAFT-DO NOT QUOTE OR CITE
-------
1 Lioy, P. J.; Wainman, T.; Zhang, J.; Goldsmith, S. (1999) Typical household vacuum cleaners: the collection
2 efficiency and emissions characteristics for fine particles. J. Air Waste Manage. Assoc. 49: 200-206.
3 Lippmann, M; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of paniculate matter components with daily
4 mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report no.
5 95.
6 Long, C. M.; Suh, H. H.; Koutrakis, P. (2000a) Characterization of indoor particle sources using continuous mass
7 and size monitors. J. Air Waste Manage. Assoc. 50: 1236-1250.
8 Long, C. M.; Suh, H. H.; Koutrakis, P. (2000b) Using time- and size-resolved paniculate data to explore infiltration
9 and deposition behavior. Environ. Sci. Technol.: accepted.
10 Lurmann, F. W.; Korc, M. E. (1994) User's guide to the regional human exposure (REHEX) model. San Francisco,
11 CA: Bay Area Air Quality Management District; draft report no. STI-93150-1414-DR.
12 Macintosh, D. L.; Xue, J.; Ozkaynak, H.; Spengler, J. D.; Ryan, P. B. (1995) A population-based exposure model
13 for benzene. J. Exposure Anal. Environ. Epidemiol. 5: 375-403.
14 Mage, D. T. (1985) Concepts of human exposure assessment for airborne paniculate matter. Environ. Int.
15 11:407-412.
16 Mage, D. T. (1998) How much PM of ambient origin are people exposed to? In: Measurement of toxic and related
17 air pollutants. Volume II: proceedings of a specialty conference; September; Gary, NC. Pittsburgh, PA: Air &
18 Waste Management Association; pp. 666-681. (A&WMA publication VIP-85).
19 Mage, D. T.; Ott, W. R. (1996) Accounting for nonuniform mixing and human exposure in indoor environments.
20 In: Tichenor, B. A., ed. Characterizing sources of indoor air pollution and related sink effects. Philadelphia,
21 PA: American Society for Testing and Materials; pp. 263-278.
22 Mage, D.; Wilson, W.; Hasselblad, V.; Grant, L. (1999) Assessment of human exposure to ambient paniculate
23 matter. J. Air Waste Manage. Assoc. 49: 174-185.
24 Mar, T. F.; Norris, G. A.; Koenig, J. Q.; Larson, T. V. (2000) Associations between air pollution and mortality in
25 Phoenix, 1995-1997. Environ. Health Perspect. 108: 347-353.
26 McBride, S. J.; Ferro, A. R.; Ott, W. R; Switzer, P.; Hildemann, L. M. (1999) Investigations of the proximity effect
27 for pollutants in the indoor environment. J. Exposure Anal. Environ. Epidemiol. 9: 602-621.
28 McCurdy, T. (1995) Estimating human exposure to selected motor vehicle pollutants using the NEM series of
29 models: lessons to be learned. J. Exposure Anal. Environ. Epidemiol. 5: 533-550.
30 McCurdy, T. (1997) Modeling the dose profile in human exposure assessments: ozone as an example. Rev. Toxicol.
31 1:3-23.
32 McCurdy, T. (2000) Conceptual basis for multi-route intake dose modeling using an energy expenditure approach.
33 J. Exposure Anal. Environ. Epidemiol. 10: 86-97.
34 McCurdy, T.; Glen, G.; Smith, L.; Lakkadi, Y. (2000) The National Exposure Research Laboratory's Consolidated
35 Human Activity Database. J. Exposure Anal. Environ. Epidemiol. 10: 566-578.
36 Mihlan, G. J.; Todd, L. A.; Truong, K. N. (2000) Assessment of occupational exposure patterns by
37 frequency-domain analysis of time series data. Appl. Occup. Environ. Hyg. 15: 120-130.
38 Miller, S. L.; Branoff, S.; Nazaroff, W. W. (1998) Exposure to toxic air contaminants in environmental tobacco
39 smoke: an assessment for California based on personal monitoring data. J. Exposure Anal. Environ.
40 Epidemiol. 8:287-311.
41 Monn, C. (2001) Exposure assessment of air pollutants: a review on spatial heterogeneity and
42 indoor/outdoor/personal exposure to suspended paniculate matter, nitrogen dioxide and ozone. Atmos.
43 Environ. 35: 1-32.
44 Monn, Ch.; Carabias, V.; Junker, M.; Waeber, R.; Karrer, M.; Wanner, H. U. (1997) Small-scale spatial variability
45 of paniculate matter <10 um (PM10) and nitrogen dioxide. Atmos. Environ. 31: 2243-2247.
46 Morawska, L.; He, C.; Hitchins, J.; Gilbert, D.; Parappukkaran, S. (2001) The relationship between indoor and
47 outdoor airborne particles in the residential environment. Atmos. Environ. 35: 3463-3473.
48 Muraleedharan, T. R.; Radojevic, M. (2000) Personal particle exposure monitoring using nephelometry during haze
49 in Brunei. Atmos. Environ. 34: 2733-2738.
50 Murray, D. M.; Burmaster, D. E. (1995) Residential air exchange rates in the United States: empirical and estimated
51 parametric distributions by season and climatic region. Risk Anal. 15: 459-465.
52 National Air Pollution Control Administration. (1969) Air quality criteria for paniculate matter. Washington, DC:
53 U.S. Department of Health, Education, and Welfare, Public Health Service; NAPCA publication no. AP-49.
54 Available from: NTIS, Springfield, VA; PB-190251/BA.
April 2002 5-115 DRAFT-DO NOT QUOTE OR CITE
-------
1 National Research Council. (1991) Human exposure assessment for airborne pollutants: advances and opportunities.
2 Washington, DC: National Academy of Sciences.
3 National Research Council. (1998) Research priorities for airborne paniculate matter. I. Immediate priorities and a
4 long-range research portfolio. Washington, DC: National Academy Press.
5 National Research Council. (1999) Research priorities for airborne paniculate matter. II. Evaluating research
6 progress and updating the portfolio. Washington, DC: National Academy Press.
7 Navidi, W.; Thomas, D.; Langholz, B.; Stram, D. (1999) Statistical methods for epidemiologic studies of the health
8 effects of air pollution. Cambridge, MA: Health Effects Institute; research report no. 86.
9 Nelson, P. R.; Conrad, F. W.; Kelly, S. P.; Maiolo, K. C.; Richardson, J. D.; Ogden, M. W. (1998) Composition of
10 environmental tobacco smoke (ETS) from international cigarettes part II: nine country follow-up. Environ.
11 Int. 24:251-257.
12 Nieuwenhuijsen, M. J.; Noderer, K. S.; Schenker, M. B.; Vallyathan, V.; Olenchock, S. (1999) Personal exposure to
13 dust, endotoxin and crystalline silica in California agriculture. Ann. Occup. Hyg. 43: 35-42.
14 Oglesby, L.; Ktinzli, N.; Roosli, M.; Braun-Fahrlander, C.; Mathys, P.; Stern, W.; Jantunen, M.; Kousa, A. (2000)
15 Validity of ambient levels of fine particles as surrogate for personal exposure to outdoor air
16 pollution—results of the European EXPOLIS-EAS study (Swiss Center Basel). J. Air Waste Manage. Assoc.
17 50: 1251-1261.
18 Ott, W. R. (1984) Exposure estimates based on computer generated activity patterns. J. Toxicol. Clin. Toxicol.
19 21:97-128.
20 Ott, W. R. (1999) Mathematical models for predicting indoor air quality from smoking activity. Environ. Health
21 Perspect. 107(suppl. 2): 375-381.
22 Ott, W.; Thomas, J.; Mage, D.; Wallace, L. (1988) Validation of the simulation of human activity and pollutant
23 exposure (SHAPE) model using paired days from the Denver, CO, carbon monoxide field study. Atmos.
24 Environ. 22:2101-2113.
25 Ott, W. R.; Mage, D. T.; Thomas, J. (1992) Comparison of microenvironmental CO concentrations in two cities for
26 human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 2: 249-267.
27 Ott, W. R.; Klepeis, N. E.; Switzer, P. (1995) Modeling environmental tobacco smoke in the home using transfer
28 functions. Presented at: 88th annual meeting and exhibition of the Air & Waste Management Association;
29 June; San Antonio, TX; paper no. 95-WP84B.03. Research Triangle Park, NC: U.S. Environmental
30 Protection Agency, Atmospheric Research And Exposure Assessment Laboratory; report no.
31 EPA/600/A-95/079. Available from: NTIS, Springfield, VA; PB95-225488.
32 Ott, W.; Switzer, P.; Robinson, J. (1996) Particle concentrations inside a tavern before and after prohibition of
33 smoking: evaluating the performance of an indoor air quality model. J. Air Waste Manage. Assoc.
34 46: 1120-1134.
35 Ott, W.; Wallace, L.; Mage, D. (2000) Predicting paniculate (PM10) personal exposure distributions using a random
36 component superposition (RCS) statistical model. J. Air Waste Manage. Assoc. 50: 1390-1406.
37 Owen, M. K.; Ensor, D. S.; Sparks, L. E. (1992) Airborne particle sizes and sources found in indoor air.
38 Atmos. Environ. Part A 26: 2149-2162.
39 Ozkaynak, H.; Spengler, J. (1996) The role of outdoor paniculate matter in assessing total human exposure.
40 In: Wilson, R.; Spengler, J., eds. Particles in our air: concentrations and health effects. Cambridge, MA:
41 Harvard University Press.
42 Ozkaynak, H.; Thurston, G. D. (1987) Associations between 1980 U.S. mortality rates and alternative measures of
43 airborne particle concentration. Risk Anal. 7: 449-461.
44 Ozkaynak, H.; Ryan, P. B.; Spengler, J. D.; Laird, N. M. (1986) Bias due to misclassification of personal exposures
45 in epidemiologic studies of indoor and outdoor air pollution. In: Berglund, B.; Berglund, U.; Lindvall, T.;
46 Spengler, J.; Sundell, J., eds. Indoor air quality: papers from the third international conference on indoor air
47 quality and climate; August 1984; Stockholm, Sweden. Environ. Int. 12: 389-393.
48 Ozkaynak, H.; Spengler, J. D.; Ludwig, J. F.; Butler, D. A.; Clayton, C. A.; Pellizzari, E.; Wiener, R. W. (1990)
49 Personal exposure to paniculate matter: findings from the Particle Total Exposure Assessment Methodology
50 (PTEAM) prepilot study. In: Indoor air '90: precedings of the 5th international conference on indoor air
51 quality and climate, volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON,
52 Canada: International Conference on Indoor Air Quality and Climate, Inc.; pp. 571-576.
53 Ozkaynak, H.; Spengler, J. D.; Xue, J.; Koutrakis, P.; Pellizzari, E. D.; Wallace, L. (1993) Sources and factors
54 influencing personal and indoor exposures to particles, elements and nicotine: findings from the particle
55 TEAM pilot study. In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma,
April 2002 5-116 DRAFT-DO NOT QUOTE OR CITE
-------
1 H., eds. Indoor air '93: proceedings of the 6th international conference on indoor air quality and climate, v. 3,
2 combustion products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93;
3 pp. 457-462.
4 Ozkaynak, H.; Xue, J.; Spengler, J.; Wallace, L.; Pellizzari, E.; Jenkins, P. (1996a) Personal exposures to airborne
5 particles and metals: results from the particle TEAM study in Riverside, California. J. Exposure Anal.
6 Environ. Epidemiol. 6: 57-78.
7 Ozkaynak, H.; Xue, J.; Weker, R.; Butler, D.; Koutrakis, P.; Spengler, J. (1996b) The particle TEAM (PTEAM)
8 study: analysis of the data: final report, volume III. Research Triangle Park, NC: U.S. Environmental
9 Protection Agency, Atmospheric Research and Exposure Assessment Laboratory; report no.
10 EPA/600/R-95/098. Available from: NTIS, Springfield, VA; PB97-102495.
11 Patterson, E.; Eatough, D. J. (2000) Indoor/outdoor relationships for ambient PM2 5 and associated pollutants:
12 epidemiological implications in Lindon, Utah. J. Air Waste Manage. Assoc. 50: 103-110.
13 Pellizzari, E. D., Mason, R. E.; Clayton, C. A.; Thomas, K. W.; Cooper, S.; Piper. L.; Rodes, C.; Goldberg, M.;
14 Roberds, J.; Michael, L. (1998) Manganese exposure study (Toronto). Final report. Research Triangle Park,
15 NC: Research Triangle Institute, Analytical and Chemical Sciences; report no. RTI/6312/02-01 F.
16 Pellizzari, E. D.; Clayton, C. A.; Rodes, C. E.; Mason, R. E.; Piper, L. L.; Fort, B.; Pfeifer, G.; Lynam, D. (1999)
17 Paniculate matter and manganese exposures in Toronto, Canada. Atmos. Environ. 33: 721-734.
18 Pellizzari, E. D.; Clayton, C. A.; Rodes, C. E.; Mason, R. E.; Piper, L. L.; Fort, B.; Pfeifer, G.; Lynam, D. (2001)
19 Paniculate matter and manganese exposures in Indianapolis, Indiana. J. Exposure Anal. Environ. Epidemiol.
20 11:423-440.
21 Phillips, K.; Howard, D. A.; Browne, D.; Lewsley, J. M. (1994) Assessment of personal exposures to environmental
22 tobacco smoke in Brttish nonsmokers. Environ. Int. 20: 693-712.
23 Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G. (1996) Assessment of air quality in Stockholm by personal
24 monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Scand. J.
25 Work Environ. Health 22(suppl. 1): 1-24.
26 Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G.; Huici, A. (1997a) Assessment of air quality in Barcelona by
27 personal monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke.
28 Environ. Int. 23: 173-196.
29 Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1997b) Assessment of air quality in Turin by personal
30 monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Environ. Int.
31 23:851-871.
32 Phillips, K.; Bentley, M. C.; Howard, D. A.; Alvan, G. (1998a) Assessment of air quality in Paris by personal
33 monitoring of nonsmokers for respirable suspended particles and environmental tobacco smoke. Environ. Int.
34 24:405-425.
35 Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1998b) Assessment of environmental tobacco smoke and
3 6 respirable suspended particle exposures for nonsmokers in Lisbon by personal monitoring. Environ. Int.
37 24:301-324.
38 Phillips, K.; Howard, D. A.; Bentley, M. C.; Alvan, G. (1999) Assessment of environmental tobacco smoke and
39 respirable suspended particle exposures for nonsmokers in Basel by personal monitoring. Atmos. Environ.
40 33: 1889-1904.
41 Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
42 (1995) Paniculate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J.
43 Respir. Crit. Care Med. 151: 669-674.
44 Pope, C. A., Ill; Hill, R. W.; Villegas, G. M. (1999) Paniculate air pollution and daily mortality on Utah's Wasatch
45 Front. Environ. Health Perspect: 107: 567-573.
46 Praml, G.; Schierl, R. (2000) Dust exposure in Munich public transportation: a comprehensive 4-year survey in
47 buses and trams. Int. Arch. Occup. Environ. Health 73: 209-214.
48 Quintana, P. J. E.; Samimi, B. S.; Kleinman, M. T.; Liu, L.-J.; Soto, K.; Warner, G. Y.; Bufalino, C.; Valencia, J.;
49 Francis, D.; Hovell, M. H.; Delfino, R. J. (2000) Evaluation of a real-time passive personal particle monitor
50 in fixed site residential indoor and ambient measurements. J. Exposure Anal. Environ. Epidemiol.
51 10:437-445.
52 Ramachandran, G.; Vincent, J. H. (1999) A Bayesian approach to retrospective exposure assessment. Appl. Occup.
53 Environ. Hyg. 14: 547-557.
April 2002 5-117 DRAFT-DO NOT QUOTE OR CITE
-------
1 Rea, A. W.; Zufall, M. I; Williams, R. W.; Sheldon, L. S.; Howard-Reed, C. (2001) The influence of human
2 activity patterns on personal PM exposure: a comparative analysis of filter-based and continuous particle
3 measurements. J. Air Waste Manage. Assoc. 51: 1271-1279.
4 Ren, P.; Jankun, T. M.; Leaderer, B. P. (1999) Comparisons of seasonal fungal prevalence in indoor and outdoor air
5 and in house dusts of dwellings in one northeast American county. J. Exposure Anal. Environ. Epidemiol.
6 9:560-568.
7 Riediker, MM.; Keller, S.; Wuthrich, B.; Koller, T.; Monn, C. (2000) Personal pollen exposure compared to
8 stationary measurements. J. Invest. Allergol. Clin. Immunol. 10: 200-203.
9 Robinson, J. P.; Wiley, J. A.; Piazza, T.; Garrett, K.; Cirksena, K. (1989) Activity patterns of California residents
10 and their implications for potential exposure to pollution. Sacramento, CA: California Air Resources Board
11 report no. CARB-A6-177-33.
12 Robinson, J. P.; Switzer, P.; Ott, W. (1994) Microenvironmental factors related to Californians' potential exposure
13 to environmental tobacco smoke (ETS). Research Triangle Park, NC: U.S. Environmental Protection Agency,
14 Atmospheric Research and Exposure Assessment Laboratory; report no. EPA/600/R-94/116. Available from:
15 NTIS, Springfield, VA; PB94-214814.
16 Rodes, C.; Sheldon, L.; Whitaker, D.; Clayton, A.; Fitzgerald, K.; Flanagan, J.; DiGenova, F.; Hering, S.;
17 Frazier, C. (1998) Measuring concentrations of selected air pollutants inside California vehicles [final
18 report]. Sacramento, CA: California Environmental Protection Agency, Air Resources Board; contract no.
19 95-339.
20 Rojas-Bracho, L.; Suh, H. H.; Koutrakis, P. (2000) Relationships among personal, indoor, and outdoor fine and
21 coarse particle concentrations for individuals with COPD. J. Exposure Anal. Environ. Epidemiol.
22 10:294-306.
23 Roorda-Knape, M. C.; Janssen, N. A. H.; De Hartog, J. J.; Van Vliet, P. H. N.; Harssema, H.; Brunekreef, B. (1998)
24 Air pollution from traffic in city districts near major motorways. Atmos. Environ. 32: 1921-1930.
25 Rotko, T.; Oglesby, L.; Ktinzli, N.; Jantunen, M. J. (2000a) Population sampling in European air pollution exposure
26 study, EXPOLIS: comparisons between the cities and representativeness of the samples. J. Exposure Anal.
27 Environ. Epidemiol. 10: 355-364.
28 Rotko, T.; Koistinen, K.; Hanninen, O.; Jantunen, M. (2000b) Sociodemographic descriptors of personal exposure
29 to fine particles (PM25) in EXPOLIS Helsinki. J. Exposure Anal. Environ. Epidemiol. 10: 385-393.
30 Ryan, P. B. (1991) An overview of human exposure modeling. J. Exposure Anal. Environ. Epidemiol. 1: 453-474.
31 Samet, J. M.; Dominici, F.; Zeger, S. L.; Schwartz, J.; Dockery, D. W. (2000) National morbidity, mortality, and air
32 pollution study. Part I: methods and methodologic issues. Cambridge, MA: Health Effects Institute; research
33 report no. 94.
34 Santos-Burgoa, C.; Rojas-Bracho, L.; Rosas-Perez, L; Ramirez-Sanchez, A.; Sanchez-Rico, G.;
3 5 Mejia-Hernandez, S. (1998) Modelaje de exposition a particulas en poblacion general y riesgo de
3 6 enfermedad respiratoria [Particle exposure modeling in the general population and risk of respiratory
37 disease]. Gac. Med. Mex. 134: 407-417.
38 Sarnat, J. A.; Koutrakis, P.; Suh, H. H. (2000) Assessing the relationship between personal paniculate and gaseous
39 exposures of senior citizens living in Baltimore, MD. J. Air Waste Manage. Assoc. 50: 1184-1198.
40 Sarnat, J. A.; Schwartz, J.; Catalano, P. J.; Suh, H. H. (2001) Gaseous pollutants in paniculate matter epidemiology:
41 confounders or surrogates? Environ. Health Perspect. 109: 1053-1061.
42 Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996) Is daily mortality associated specifically with fine particles? J.
43 Air Waste Manage. Assoc. 46: 927-939.
44 Sexton, K.; Ryan, P. B. (1988) Assessment of human exposure to air pollution: methods, measurements, and
45 models. In: Watson, A. Y.; Bates, R. R.; Kennedy, D., eds. Air pollution, the automobile, and public health.
46 Washington, DC: National Academy Press; pp. 207-238.
47 Sexton, K.; Spengler, J. D.; Treitman, R. D. (1984) Personal exposure to respirable particles: a case study in
48 Waterbury, Vermont. Atmos. Environ. 18: 1385-1398.
49 Sheldon, L.; Clayton, A.; Jones, B.; Keever, J.; Perritt, R.; Smith, D.; Whitaker, D.; Whitmore, R. (1992a) Indoor
50 pollutant concentrations and exposures. Sacramento, CA: California Air Resources Board; report no.
51 ARB/R-92/486.
52 Sheldon, L.; Clayton, A.; Keever, J. Perritt, R.; Whitaker, D. (1992b) PTEAM: monitoring of phthalates and PAHs
53 in indoor and outdoor air samples in Riverside, California, volume II [final report]. Sacramento, CA:
54 California Environmental Protection Agency, State Air Resources Board, Research Division; contract no.
55 A933-144. Available from: NTIS, Springfield, VA; PB93-205649/XAB.
April 2002 5-118 DRAFT-DO NOT QUOTE OR CITE
-------
1 Sheldon, L.; Whitaker, D.; Keever, I; Clayton, A.; Perritt, R. (1993a) Phthalates and PAHs in indoor and outdoor
2 air in a southern California community. In: Jantunen, M; Kalliokoski, P.; Kukkonen, E.; Saarela, K.;
3 Seppanen, O.; Vuorelma, H., eds. Indoor air '93: proceedings of the 6th international conference on indoor
4 air quality and climate. Volume 3: combustion products, risk assessment, policies; July; Helsinki, Finland.
5 Helsinki, Finland: Indoor Air '93; pp. 109-114.
6 Sheldon, L.; Clayton, A.; Keever, J.; Perritt, R.; Whitaker, D. (1993b) Indoor concentrations of polycyclic aromatic
7 hydrocarbons in California residences. Sacramento, CA: Air Resources Board; ARE contract no. A033-132.
8 Sheldon, L.; Clayton, A.; Perritt, R.; Whitaker, D. A.; Keever, J. (1993c) Indoor concentrations of polycyclic
9 aromatic hydrocarbons in California residences and their relationship to combustion source use. In: Jantunen,
10 M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air '93: proceedings
11 of the 6th international conference on indoor air quality and climate. Volume 3: combustion products, risk
12 assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 29-34.
13 Shy, C. M.; Kleinbaum, D. G.; Morgenstern, H. (1978) The effect of misclassification of exposure status in
14 epidemiological studies of air pollution health effects. Bull. N. Y. Acad. Med. 54: 1155-1165.
15 Siegmann, K.; Scherrer, L.; Siegmann, H. C. (1999) Physical and chemical properties of airborne nanoscale
16 particles and how to measure the impact on human health. J. Mol. Struct. (Theochem) 458: 191-201.
17 Spengler, J. D.; Treitman, R. D.; Tosteson, T. D.; Mage, D. T.; Soczek, M. L. (1985) Personal exposures to
18 respirable particulates and implications for air pollution epidemiology. Environ. Sci. Technol. 19: 700-707.
19 Spengler, J. D.; Ozkaynak, H.; Ludwig, J.; Allen, G.; Pellizzari, E. D.; Wiener, R. (1989) Personal exposures to
20 paniculate matter: instruments and methodologies P-TEAM. In: Measurement of toxic and related air
21 pollutants: proceedings of the 1989 U.S. EPA/A&WMA international symposium; May; Raleigh, NC.
22 Pittsburgh, PA: Air & Waste Management Association; pp. 449-463. (A&WMA publication VTP-13).
23 Spier, C. E.; Little, D. E.; Trim, S. C.; Johnson, T. R.; Linn, W. S.; Hackney, J. D. (1992) Activity patterns in
24 elementary and high school students exposed to oxidant pollution. J. Exposure Anal. Environ. Epidemiol.
25 2:277-293.
26 Suggs, J. C.; Burton, R. M. (1983) Spatial characteristics of inhalable particles in the Philadelphia metropolitan
27 area. J. AirPollut. Control Assoc. 33: 688-691.
28 Suh, H. H.; Spengler, J. D.; Koutrakis, P. (1992) Personal exposures to acid aerosols and ammonia. Environ. Sci.
29 Technol. 26:2507-2517.
30 Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993a) Indoor and outdoor acid aerosol and gas concentrations.
31 In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air
32 '93: proceedings of the 6th international conference on indoor air quality and climate. Volume 3: combustion
33 products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 257-262.
34 Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1993b) Validation of personal exposure models for sulfate and aerosol
35 strong acidity. J. Air Waste Manage. Assoc. 43: 845-850.
36 Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1994) The relationship between airborne acidity and ammonia in indoor
37 environments. J. Exposure Anal. Environ. Epidemiol. 4: 1-23.
38 Suh, H. H.; Allen, G. A.; Koutrakis, P.; Burton, R. M. (1995) Spatial variation in acidic sulfate and ammonia
39 concentrations within metropolitan Philadelphia. J. Air Waste Manage. Assoc. 45: 442-452.
40 Tamura, K.; Ando, M.; Sagai, M.; Matsumoto, Y. (1996a) Estimation of levels of personal exposure to suspended
41 paniculate matter and nitrogen dioxide in Tokyo. Environ. Sci. (Tokyo) 4: 37-51.
42 Tamura, K.; Ando, M.; Miyazaki, T.; Funasaka, K.; Kuroda, K. (1996b) Indoor concentrations and personal
43 exposure to paniculate matter at roadside houses. In: Indoor air '96: the 7th international conference on
44 indoor air quality, volume I; July; Nagoya, Japan. Tokyo, Japan: Institute of Public Health; pp. 599-602.
45 Teschke, K.; Marion, S. A.; Vaughan, T. L.; Morgan, M. S.; Camp, J. (1999) Exposures to wood dust in U.S.
46 industries and occupations, 1979 to 1997. Am. J. Ind. Med. 35: 581-589.
47 Thomas, K. W.; Pellizzari, E. D.; Clayton, C. A.; Whitaker, D. A.; Shores, R. C.; Spengler, J.; Ozkaynak, H.;
48 Froehlich, S. E.; Wallace, L. A. (1993) Particle total exposure assessment methodology (PTEAM) 1990
49 study: method performance and data quality for personal, indoor, and outdoor monitoring. J. Exposure Anal.
50 Environ. Epidemiol. 3:203-226.
51 Tsang, A. M.; Klepeis, N. E. (1996) Descriptive statistics tables from a detailed analysis of the national human
52 activity pattern survey (NHAPS) data. Las Vegas, NV: U.S. Environmental Protection Agency, Office of
53 Research and Development; report no. EPA/600/R-96/148.
April 2002 5-119 DRAFT-DO NOT QUOTE OR CITE
-------
1 U.S. Environmental Protection Agency. (1982) Air quality criteria for paniculate matter and sulfur oxides. Research
2 Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
3 Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
4 U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
5 NC: National Center for Environmental Assessment-RTF Office; report nos. EPA/600/P-95/00laF-cF. 3v.
6 U.S. Environmental Protection Agency. (1998) Paniculate matter research needs for human health risk assessment
7 to support future reviews of the national ambient air quality standards for paniculate matter. Research
8 Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-97/132F.
9 U.S. Environmental Protection Agency. (2000) Aerometric Information Retrieval System (AIRS). Research
10 Triangle Park, NC: Office of Air Quality Planning and Standards. Available: www.epa.gov/ttn/airs/
11 [2000, October 6].
12 Vakeva, M; Hameri, K.; Kulmala, M; Lahdes, R.; Ruuskanen, J.; Laitinen, T. (1999) Street level versus rooftop
13 concentrations of submicron aerosol particles and gaseous pollutants in an urban street canyon.
14 Atmos. Environ. 33: 1385-1397.
15 Van Vliet, P.; Knape, M.; De Hartog, J.; Janssen, N.; Harssema, H.; Brunekreef, B. (1997) Motor vehicle exhaust
16 and chronic respiratory symptoms in children living near freeways. Environ. Res. 74: 122-132.
17 Vette, A. F.; Rea, A. W.; Lawless, P. A.; Rodes, C. E.; Evans, G.; Highsmith, V. R.; Sheldon, L. (2001)
18 Characterization of indoor-outdoor aerosol concentration relationships during the Fresno PM exposure
19 studies. Aerosol Sci. Technol.: accepted.
20 Vyas, V. M.; Sun, Q.; Chandrasekar, A.; Wang, S.; Georgopoulos, P. G.; Burke, J.; Xue, J.; Ozkaynak, H. (2002)
21 Source-to-dose analysis of population exposures to PM. Presented at: Eleventh annual meeting of the
22 International Society of Exposure Analysis; November 2001; Charleston, SC.
23 Wainman, T.; Zhang, J.; Weschler, C. J.; Lioy, P. J. (2000) Ozone and limonene in indoor air: a source of
24 submicron particle exposure. Environ. Health Perspect. 108: 1139-1145.
25 Waldman, J. M.; Liang, C. S. K. (1993) Penetration and neutralization of acid aerosols at institutional settings.
26 In: Jantunen, M.; Kalliokoski, P.; Kukkonen, E.; Saarela, K.; Seppanen, O.; Vuorelma, H., eds. Indoor air
27 '93: proceedings of the 6th international conference on indoor air quality and climate. Volume 3: combustion
28 products, risk assessment, policies; July; Helsinki, Finland. Helsinki, Finland: Indoor Air '93; pp. 263-268.
29 Waldman, J. M.; Liang, C. S. K.; Menon, P. (1990) Indoor exposures to acidic aerosols at child and elderly care
30 facilities. In: Indoor air '90: precedings of the 5th international conference on indoor air quality and climate,
31 volume 2, characteristics of indoor air; July-August; Toronto, ON, Canada. Ottawa, ON, Canada:
32 International Conference on Indoor Air Quality and Climate, Inc.; pp. 607-612.
33 Waldman, J. M.; Lioy, P. J.; Greenberg, A.; Butler, J. P. (1991) Analysis of human exposure to benzo(a)pyrene via
34 inhalation and food ingestion in the Total Human Environmental Exposure Study (THEES). J. Exposure
35 Anal. Environ. Epidemiol. 1: 193-225.
36 Wallace, L. (1996) Indoor particles: a review. J. Air Waste Manage. Assoc. 46: 98-126.
37 Wallace, L. (2000a) Real-time monitoring of particles, PAH, and CO in an occupied townhouse. Appl. Occup.
38 Environ. Hyg. 15: 39-47.
39 Wallace, L. (2000b) Correlations of personal exposure to particles with outdoor air measurements: a review of
40 recent studies. Aerosol Sci. Technol. 32: 15-25.
41 Wallace, L.; Quackenboss, J.; Rodes, C. (1997) Continuous measurements of particles, PAH, and CO in an
42 occupied townhouse in Reston, VA. In: Measurement of toxic and related air pollutants. Volume II:
43 proceedings of a specialty conference; April-May; Research Triangle Park, NC. Pittsburgh, PA: Air & Waste
44 Management Association, pp. 860-871. (A&WMA publication VIP-74).
45 Watson, J. G.; Cooper, J. A.; Huntzicker, J. J. (1984) The effective variance weighting for least squares calculations
46 applied to the mass balance receptor model. Atmos. Environ. 18: 1347-1355.
47 Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen, Q.
48 (1990) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
49 Watson, J. G.; Chow, J. C.; Pace, T. G. (1991) Chemical mass balance. In: Hopke, P. K., ed. Receptor modeling for
50 air quality management. New York, NY: Elsevier; pp. 83-116. (Data handling in science and technology—
51 volume 7).
52 Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
53 Environ. 33:2301-2312.
April 2002 5-120 DRAFT-DO NOT QUOTE OR CITE
-------
1 Wheeler, A. J.; Williams, I.; Beaumont, R. A.; Hamilton, R. S. (2000) Characterisation of paniculate matter
2 sampled during a study of children's personal exposure to airborne paniculate matter in a UK urban
3 environment. Environ. Monit. Assess. 65: 69-11.
4 White, W. H. (1998) Statistical considerations in the interpretation of size-resolved paniculate mass data. J. Air
5 Waste Manage. Assoc. 48: 454-458.
6 Wiener, R. W. (1988) Measurement and evaluation of personal exposure to aerosols. In: Measurement of toxic and
7 related air pollutants: proceedings of the 1988 EPA/APCA international symposium; May; Research Triangle
8 Park, NC. Pittsburgh, PA: Air Pollution Control Association; pp. 84-88. (APCA publication VIP-10).
9 Wiener, R. W. (1989) Particle total exposure assessment methodology~an overview of planning and
10 accomplishments. In: Measurement of toxic and related air pollutants: proceedings of the 1989 U.S.
11 EPA/A&WMA international symposium; May; Raleigh, NC. Pittsburgh, PA: Air & Waste Management
12 Association; pp. 442-448. (A&WMA publication VIP-13).
13 Wiener, R. W.; Wallace, L.; Pahl, D.; Pellizzari, E.; Whittaker, D.; Spengler, J.; Ozkaynak, H. (1990) Review of the
14 particle TEAM 9 home field study. In: Measurement of toxic and related air pollutants: proceedings of the
15 1990 EPA/A&WMA international symposium, v. 2; May; Raleigh, NC. Pittsburgh, PA: Air & Waste
16 Management Association; pp. 452-460. (A&WMA publication VIP-17).
17 Wiley, J. A.; Robinson, J. P.; Piazza, T.; Garrett, K.; Cirksena, K.; Cheng, Y.-T.; Martin, G. (1991a) Activity
18 patterns of California residents. Final report. Sacramento, CA: California Air Resources Board; report no.
19 ARB/R93/487. Available from: NTIS, Springfield, VA.; PB94-108719.
20 Wiley, J. A.; Robinson, J. P.; Cheng, Y.-T.; Piazza, T.; Stork, L.; Pladsen, K. (1991b) Study of children's activity
21 patterns: final report. Sacramento, CA: California Air Resources Board; report no. ARB-R-93/489.
22 Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
23 (2000a) The 1998 Baltimore Paniculate Matter Epidemiology-Exposure Study: part 1. Comparison of
24 ambient, residential outdoor, indoor and apartment paniculate matter monitoring. J. Exposure Anal. Environ.
25 Epidemiol. 10: 518-532.
26 Williams, R.; Suggs, J.; Creason, J.; Rodes, C.; Lawless, P.; Kwok, R.; Zweidinger, R.; Sheldon, L. (2000b) The
27 1998 Baltimore paniculate matter epidemiology-exposure study: part 2. Personal exposure assessment
28 associated with an elderly study population. J. Exposure Anal. Environ. Epidemiol.: 10: 533-543.
29 Williams, R.; Creason, J.; Zweidinger, R.; Watts, R.; Sheldon, L.; Shy, C. (2000c) Indoor, outdoor, and personal
3 0 exposure monitoring of paniculate air pollution: the Baltimore elderly epidemiology-exposure pilot study.
31 Atmos. Environ. 34: 4193-4204.
32 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
33 epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
34 Wilson, W. E.; Mage, D. T.; Grant, L. D. (2000) Estimating separately personal exposure to ambient and
3 5 nonambient paniculate matter for epidemiology and risk assessment: why and how. J. Air Waste Manage.
36 Assoc. 50: 1167-1183.
37 Yakovleva, E.; Hopke, P. K.; Wallace, L. (1999) Receptor modeling assessment of particle total exposure
38 assessment methodology data. Environ. Sci. Technol. 33: 3645-3652.
39 Zeger, S. L.; Thomas, D.; Dominici, F.; Samet, J. M.; Schwartz, J.; Dockery, D.; Cohen, A. (2000) Exposure
40 measurement error in time-series studies of air pollution: concepts and consequences. Environ. Health
41 Perspect. 108: 419-426.
42 Zmirou, D.; Masclet, P.; Boudet, C.; Dor, F.; Dechenaux, J. (2000) Personal exposure to atmospheric polycyclic
43 aromatic hydrocarbons in a general adult population and lung cancer risk assessment. J. Occup. Environ.
44 Med. 42: 121-126.
April 2002 5-121 DRAFT-DO NOT QUOTE OR CITE
------- |